Exploring portable ratings for gig workers
medium.com
medium.com
Gig workers who find themselves in a rut of poor performance for one reason or another just aren't going to be able to escape the old judgements of their performance for much longer periods of time. It's oppressive.
Making reputation-follow "opt-in" makes a lot of sense. (Perhaps with egregiously negative reputations being weighted at least somewhat.)
It discourages scammers because customers know even workers that are new to the platform have built up a good reputation elsewhere, and they know that anything seriously bad will follow them.
I see primary difficulty in how to communicate the "egregiously bad reviews will follow you" effect and how to determine which reviews are egregious. Plus it's like 100% probability of it being gamed with bad-actors somehow...
(Personally I'd define the "egregiousness scale" to be a time-weighted average that's also weighted by the percentile of the reviewer's average rating - if you give all your gig workers 3 stars, then 3 stars is effectively 5 stars for you so workers see that benefit, and reviews further in the past count less against you both pro and con. But I'm not qualified to know if this would actually work out.)
If there is a new gig platform and I have 0 rep, that's fine because everyone has 0 rep.
If there is a new gig platform and I have 0 rep, and a lot of people brought their portable reputation with them, it means I have a disadvantage.
It does provide more knowledge to consumers, and it does allow people to make different choices about who works for them.
Gig marketplaces come and go. If you spend 5 years working dozens of jobs a week to build up thousands of positive reviews only to have the marketplace go belly-up and now you have to start at zero, how do you think the gig economy can retain actual talent that expects to get paid well for their proven track-record?
Or put it another way: I personally like the option of being able to pay more for people with a long history of positive reviews. I'm willing to pay much less to hire an untrusted worker. I don't think I'm alone. If everything starts at zero every few years, then all the wages reset to baseline every few years as well.
Sure let people start at zero, but also let them get paid like an untrusted worker. And sure, let people bring over their experience. But let them get paid and charge like they have that experience.
Every policy has winners and losers. I just felt like on this discussion thread there was a loser (people entering the gig economy) not mentioned.
While the first problem is obvious, the latter is especially hard and I have yet to see it solved properly. The only place that comes to my mind that has somewhat mitigated it was early eBay that had ratings for both sellers and clients, so dishonest client would receive low ratings. But it also tended to discriminate new users as there's no way to distinguish between a honest, new user and just another single-use account for a scam. Furthermore, it only works on an established system, as you need to have many ratings to filter out noise, so it's a chicken/egg game ("nobody uses platform, because there's not enough users" / "there's not enough users, because nobody uses the platform").
I don't see how these problems could be solved without a 3rd party arbitration, but to do this you have to move back to square 1 (centralized service).
You’re also forgetting sabotage by competitors.
So in China they're looking at a Social Credit System (which has been talked about here before [1][2]) which is a government program which would apply to all Chinese citizens and rate their trustworthiness. It's seen as authoritarian, providing no way to start over, etc.
Here in America we've got this post and it's implication that we're going to institute a similar social credit system except it will come from corporations, not government and it won't apply to everyone, only gig workers. It's seen as an interesting technical problem and a way to increase market efficiency.
A tale of two viewpoints.
...criticism in this very thread they almost exactly mirrors the criticism you point to by Americans of the Chinese system, particularly on the “no way to start over”.
China: Manage to upset, doesn't matter how, the party representative and you're doomed. West: s/party/company/
Neither approach brings truth, just verisimilitude.
Every review or score out there has been gamed in some way either postively by friends and colleagues or negatively by your competitors.
Now, as an insurer, you have a monetary interest in ending up with as few riskier clients as you can. You can try to use pricing or exclusions to try to achieve this, but regulations often step in, because otherwise only the least risky clients could obtain insurance, and everyone else would have to fend for themselves. However, the least risky people often don't even want to purchase insurance, because they perceive its utility to be low, and the incidence of catastrophic events to be unlikely.
This generally means that clients who perceive themselves to be low-risk will place a lower value on insurance, necessitating a lower price, or they'll take their business elsewhere or nowhere at all. People who perceive themselves to be higher-risk will "accept" a higher price, because they expect to have a better outcome with insurance than without. Since companies want to avoid this type of customer, prices can creep higher than they need to be just by factoring in their risk.
Pretending you don't know the risk factor when you do (perhaps in the name of fairness) can also cause problems.
I don't think they should be allowed to use traits that you can't change like age, sex, and race in the same fashion as employment where you're not allowed to discriminate against women of child bearing age even if it would be economical.
The Black Mirror was likely inspired by the Chinese system.
Gig workers would be able to avoid the consequences of their rating for everything but their gig jobs.
The major difference is who owns the rating system and which factors modify the rating.
Just wait for those networks to eventually pull a yelp on their members: "nice five star ratings you got there, would be a shame if something... happened to them. My brother, well, he can be a little clumsy at times when importing a migrated profile. Things get ...lost. Please contact us for a quote for our Premium Verified Trust package".
When acting on behalf of shareholders, every scrap of money left on the table is a betrayal.
I think your links to China's "social credit" system have invited comments that derailed what this thread's submitted article is actually about.[1]
They are different situations and motivations.
The Chinese social rating is imposed from the top down on its citizens. It's possible most of them don't want it.
On the other hand, this particular article is about about tradespeople (think carpenters, freelance programmers, etc) that want to have a reputation -- so they are more marketable to potential clients. The problem for them is that their reputation karma is locked and controlled in a proprietary platform (e.g. Upwork, or AngiesList, etc). It's not portable.
The article discusses how tradespeople are unhappy that they can't get references from previous projects or a gig platform deletes their ratings.) She's saying there's a bottom up set of voluntary market desires and wondering if a technical solution can be developed for them. From those pain points, the author hypothesizes some ideas about decentralizated ownership of ratings.
Yes, the China ratings system is also an important human rights issue but that's a distraction from this particular article. They are not the same.
[1] The main point of the article starts halfway down with: "It was a radically different story for tradespeople. Their reputation data feels important to them, and they prefer to keep control over it. They preferred word of mouth reputation and recommendations, as there was no middleman who could take that away from them. Online platforms were seen as something to graduate away from once you had a sufficient “real world” presence. [...]"
>On the other hand, this particular article is about about tradespeople (think carpenters, programmers, etc) that want to have a reputation -- so they are more marketable to potential clients. The problem for them is that their reputation karma is locked and controlled in a proprietary platform (e.g. Upwork, or AngiesList, etc). It's not portable.
Hmm, well let's compare.
From this article:
'I tried to reach them and I was really upset because I had built that profile... I couldn't take my profile and move it. I lost everything but they didn't really care because it says in their terms and conditions they can close any account without reason. -Nick
From this[1] article:
Chen Chao, a 34-year-old vendor also waiting in line, agreed. “If you didn’t commit a crime, why do you need to look at your file?” he asked, adding that he’s looking forward to an e-system. “An electric file will be more convenient for me. I believe the workers here have professional ethics, so they won’t leak my information.”
and
“I just opened the app, showed my score, they took down my name and phone number and I breezed through in five minutes,” said Yolanda Liu, 30, who works for a state-owned sports organisation. “China needs a credit system so that people like me who are responsible can get more benefits.”
Is China really imposed? After all look! People there really want to have a reputation - so they can get credit advantages. The problem for them is that their reputation karma is locked and controlled on the dang'an paper record system.
>The article discusses how tradespeople are unhappy they can't get references from previous jobs or a gig platform deletes their ratings.) From those pain points, the author hypothesizes some ideas about decentralization of ratings.
That SCMP article discusses how Chinese citizens are unhappy they can't get advantages for being a good citizen. From those pain points, the Chinese government has hypothesized some ideas about a rating system. Is your only problem with the Chinese system the centralization?
>She's saying there's a bottom up set of voluntary market desires and wondering if a technical solution can be developed for them.
And when it comes out are you going to use a service from gig workers who opt out? If American gig workers' livelihoods depend on using portable ratings, maybe you should reconsider how voluntary it really is.
>Yes, the China ratings system is also an important human rights issue
I'm curious, why do you think it's a human rights issue? Is it just that it's controlled by the government? I wouldn't think that would, in itself, constitute a human rights issue so there must be another component.
>I think your links to China's "social credit" system have invited comments that have derailed what this thread's submitted article is actually about
Well no, I think my links to China's social credit system have shown what 'portable ratings' really are. Truth through analogy maybe.
[1] http://www.scmp.com/news/china/policies-politics/article/188...
Yes. The official Chinese government is developing it and planning to force it on its citizens. The soundbites from a Chinese citizen that wants it may not be representative of the China population.
On the other hand, this article is spitballing some technical ideas to help freelancers. The author, Cassie Robinson is not a government official forcing these ratings on everyone. Ultimately, her idea may (1) gain zero traction and (2) be technically unfeasible.
>I'm curious, why do you think it's a human rights issue?
The common complaints center on "right to be forgotten", "unfairness", "ostracization", etc.
>And when it comes out are you going to use a service from gig workers who opt out?
I think many people will. Today, many hire contractors that don't have ratings on AngiesList. Not sure why you think the only end game is that people who opt out of portable ratings will get no work.
So your only problem is that it's the government? I guess I see a lot more wrong with it. Actually that's unfair you do see another problem with it:
>The common complaints center on "right to be forgotten", "unfairness", "ostracization", etc.
So don't American workers deserve those same rights? How is it that an Uber Driver or Food-Delivery Professional or whatever we call it has a "right to be forgotten" when the only way for him to get work is to consent to a permanent report card which follows him wherever he goes?
>some technical ideas to help freelancers.
That's exactly my point on viewpoint! Right, this is an 'interesting technical problem' while Social Credit is 'Orwellian.'
>On the other hand, this article is spitballing
People in the Chinese government were just spitballing at one point as well. I'm just pointing and saying: "that's a nasty looking spitball Cassie just hocked up." Besides she mentions two companies working in this space, quick tell me, doesn't this sound like Chinese Social Credit?
"Empowering trustworthy people to build a fairer world."
"the old credit score system does not recognize you as a trustworthy citizen. Prove how wrong they are and get the conditions you deserve."
"We need a new form of credit score that empowers people. Fortunately today we have new data sources to assess whether people behave well and are trustworthy. Traity´s algorithms use non-credit data like eBay or AirBnB profiles, FB and LinkedIn social networks, to determine people’s trustworthiness, giving you the chance to get access to the services you need at the prices you deserve."
Yeah, they're all from traity.com. This is something that we need to push back against.
>The author, Cassie Robinson is not a government official forcing these ratings on everyone.
Correct.
Those rights are relevant to China's forced social credit but are not relevant to the decentralized ratings _owned_ by freelancers. Your doomsday scenario of "only way to get work is permanent ratings" is hyperbolic.
Consider a freelance camera man, or freelance carpenter. At the JohnDoeCameraMan.com webpage, he has:
+ customer testimonials
+ customer references
+ endorsements
+ portfolio of previous work from past customers (film credits, film stills, etc)
These 4 "ratings", are already decentralized and owned on his webpage. He also wants to own "platform" ratings as well.
Why are voluntary decentralized ratings a scary bogeyman that threatens other freelancers with joblessness? You say that some freelancers that use portable ratings to their advantage pressures other freelancers to carry ratings because that's the "only way to get work."
But people have been getting work without a history of customer testimonials, and without references, without endorsements, without AngiesList ratings, etc.
I'm not convinced that an optional portable rating system creates a defacto forced system. None of the previous social signals of past work history has created a defacto forced itemization of customer references & testimonials.
>tell me, doesn't this sound like Chinese Social Credit?
No, the China government controls the ratings by a central authority. It's very easy to see how it can be used as a tool for opppression.
I don't agree that there's a equivalent leap between freelancers owning their ratings to the China system.
>This is something that we need to push back against.
Who is "we"? Are you speaking for the freelancers who _want_ their portable ratings?
We seem to be coming at this from very different philosophical positions. Let me ask an honest question to get a baseline of your thinking...
Do you believe a carpenter should delete "customer testimonials and references" from his website because if he leaves it there, it makes it unfair for the other carpenters who don't have customer testimonials?
If yes, we are _way_ apart on seeing how society wants to naturally interact. We certainly wouldn't be able to bridge that massive canyon by exchanging HN comments.
Consider a freelance camera man, or freelance carpenter. At the JohnDoeCameraMan.com webpage, he has:
+ customer testimonials
+ customer references
+ endorsements
+ portfolio of previous work from past customers (film credits, film stills, etc)
Exactly! Because you're not going to use the freelance carpenter that doesn't have at least a few of those things right? In fact, basically (or completely) no one will. So the freelance carpenter is compelled by the market to offer at least a few of those things on his website. I wouldn't call that voluntary because his livelihood depends on having a few testimonials or endorsements. To him it's exactly the same whether individuals in the market refuse to buy his service because he has no testimonials or whether the government tells him he can't sell his service because he has no testimonials. Both situations are equally voluntary (which is to say not at all.)
Now if you had a much more objective measure out there, a lifetime score where you can never hide any blemish, which is systematic across time you'd be much more likely to use that right? After all testimonials, good experiences, bad experiences, it's all in there. The basic problem here is that the market is going to prefer a lifetime unscrubbable rating history. It's straight up more information. So carpenters and camera men will be coerced into this system - the same one that you believe violates human rights.
>I'm not convinced that an optional portable rating system creates a defacto forced system. None of the previous social signals of past work history has created a defacto forced itemization of customer references & testimonials.
Yes, you don't have to itemize it any specific way, but you still have to have them and that's the point.
>Who is "we"? Are you speaking for the freelancers who _want_ their portable ratings?
Do you think the Chinese should stand against Social Credit? What about the ones who want it?
>Do you believe a carpenter should delete "customer testimonials and references" from his website because if he leaves it there, it makes it unfair for the other carpenters who don't have customer testimonials?
No because testimonials aren't held across all time, in the current system I can start over.
My problem with both rating systems is twofold. (1) the idea that it erases a right to be forgotten and (2) that people will be coerced into the system.
Because the "want" is coupled with the involuntary conscription of those who don't want it. That's not just an academic difference. That some in China "want" it is almost irrelevant since China's government is implementing it anyway.
>you're not going to use the freelance carpenter that doesn't have at least a few of those things right? In fact, basically (or completely) no one will.
But like I tried to emphasize earlier, this does _not_ actually happen in reality. People still get hired all the time _without_ those customer histories. It truly is optional.
>livelihood depends on having a few testimonials or endorsements. [...] , but you still have to have them and that's the point.
And my point is that you do not have to have them. Otherwise, you'd have an infinite regress contradiction of nobody being hired because every carpenter at some point had zero customer history of testimonials. Carpenters eventually grow old and die which would leave us with new carpenters with no customer history. Nobody hires them, and therefore, nothing gets built. Obviously, if that apocalypse hasn't happened, it means real customers do in fact constantly hire carpenters without any testimonials. (Heck, most carpenters probably don't even have websites.)
In other words, the subset of carpenters that went the extra mile to place testimonials on their webpage did not drive other carpenters out of work.
Likewise, people might hypothesize that influential contractor websites like AngiesList would make contractors _not_ on that platform unable to get jobs. But their prediction would be wrong because that didn't happen. Lots of contractors ignore that platform even though AngiesList has had 22 years of existence and market penetration.
Same optionality will happen with freelancer ratings.
The China system frightens me. The freelancer ratings do not.
If anything, the more likely scenario is that the decentralized freelancer ratings will be ignored because too many bad actors will figure out how to create fake ratings from fake customers. Having a decentralized rating might be even be a negative signal: Customers would think that "only the scammers" would bother to prominently display their decentralized freelancer rating. This is similar to BBB Better Business Bureau "A" rating prominently displayed by many dubious and unethical companies. The BBB "seal of approval" becomes a joke.
The China social credit scoring is harder to ignore because it's centrally controlled and administered and everybody is involuntarily included instead of a voluntary subset like gig contractors. This makes a profound difference.
>No because testimonials aren't held across all time,
Does that mean you would be against contractors placing positive customer testimonials into an immutable blockchain?
>My problem with both rating systems is twofold. (1) the idea that it erases a right to be forgotten
I don't know all the details but presumably, the decentralized ratings would be an opt-in system. If a contractor wants to be "forgotten", just ignore it and don't create a profile on it. Same as if he ignores other decentralized systems such as Diaspora.
>and (2) that people will be coerced into the system.
I see no evidence that this will happen for contractors. History has shown that some contractors do "extra credit" to make them stand above the crowd and yet the majority of their peers will still not be (socially) pressured to emulate them. I also see no evidence that buyers of freelance work (such as homeowners en masse) converge to an ultimatum of only hiring contractors that emulate the contractors that advertise their customer histories.
But it wasn't an issue for the "gig economy" platforms. They are happy to lock in workers by not exporting ratings. On top of that, accepting imported ratings would provide a boost to the competitors that export ratings. So, what's the business case?
Pretty design, but I'm amazed somebody could write an article covering so many detailed questions and not even touch on this rather fundamental one.
It reminds me a bit of the heady OpenID days where every piddling website was an OpenID IdP (identity provider), but next to no websites actually accepted OpenID logins.
The problem is that when something as important as someone's livelihood depends on factors outside their control (like someone else's subjective impression of their performance) then it opens the system up to corruption/gaming/extortion.
It's probably valid to keep a log of someone's employee history. But it must be under that person's control and they must have avenues for disputing its contents or expunging it in its entirety (similar to a bankruptcy).
If we go down this road, then it must be balanced by employer ratings such as payment followthrough.
This is a shorthand way of saying the drivers themselves aren't even checking their ratings, as long as they're above the performance threshold! This paragraph gives away who 'ratings portability' is really serving, and it is not the drivers.
A better system would be to just force people to wear their overall positive rating percentage visibly on their clothes in large red print. We can call it... the scarlet number.
Online services have been around for a long time and yet very few implement data portability. Often, it's part of their moat against competitors and ensures lock-in. Few incentives exist for incumbents to offer data export.
From a social standpoint, this raises some serious issues. Distributed reputation systems exist informally, whose lack of universal applicability give them some protection against trivial malice. In the financial world, various credit scoring systems use trusted third parties to aggregate information from entities one has transacted with, and offer a service for other parties to obtain this information on the user's request. In some places government-issued strong authentication is used to establish identity, but in the US, weak identification and publicly available information is used as a challenge to the user to establish their identity. Both of them come with a host of issues that ought to be cause for concern.
If no link to some "higher form" of identity is implied, then the problem becomes almost trivial: each user of 'Service A' can become a cross-referenced user on 'Service B' by proving control over their account in 'Service A'. But this too is vulnerable to credential theft, and new guises can be created at any time by just creating additional accounts in 'Service A'.
[0] For example.