I am not defending Shipt and there is no doubt gig workers are in a very vulnerable position. However, the data analysis results as presented in the article do not support the article's main point. "40% are getting paid at least 10% less" is not unnatural to expect whenever pay is redistributed, especially since some 30+% are getting at least 10% more. Imagine a _hypothetical_ situation where Shipt is 100% on point and driving a fairer version of the algorithm patch removing a way for workers to "optimize" for short, well paid trips, resulting in pay cuts to those who had learnt how to do it, while not changing/increasing pay for everyone else. We would see the same kind of result: some portion of workers would get paid 10% less, some 10% more. This does show that workers are paid differently for the same work they have been doing, but does not prove the change is unfair.
I had similar thoughts. But let's not overlook the information asymmetry, which contributed to the dissatisfaction. I don't want to live in a world which is controlled unilaterally, and intransparently by a group of people who assume they have a full picture of the situation and assume they understand moral completely, and also don't think it necessary to explain how they think so highly of themselves.
It's an interesting question, as we have a spectrum from 'little to no transparency' through to 'full transparency' (which is pretty rare), and in the middle sits the usual approach of 'communications-team-led messaged quasi-transparency'. Difficult to know (without more info) where Shipt would have appeared on this spectrum, but given the issue, they're probably somewhere towards the 'insufficient transparency' end.
What's silly in this case is that (as others have pointed out) the new algorithm seems to have been reasonably equitable, with a genuine redistribution of payments, rather than just a cut overall. Shipt could have avoided this whole situation with a straightforward explanation of the changes, together with a few examples of the cases/jobs in which people would earn more or less.
I think the issue is that there was full transparency on pay (a fixed base rate plus a fixed percentage) and then it was changed without warning.
I work for a salary, which is fully transparent in the sense that I know what my next paycheck will be to the penny. (It’s not transparent in how it’s set, but it is week-to-week.) If my employer started paying me based on effort, and didn’t tell me what constituted effort, not only would I be pissed off but that would be completely illegal.
I’m not suggesting that this change is or should be illegal. But if it happened to me I’d find it extremely unfair.
If Shipt is actually trying to incentivize better performance, it seems the best way is to be completely transparent about the rewards algorithm. "Short high-value trips are now somewhat de-rated, and trips requiring more effort now have improved rewards, specifically ..." or whatever.
This "communications team" approach did everyone a disservice if Shipt mgt were really trying to improve results.
OTOH, if the actual goal was to screw workers harder, they accomplished that, as here ate arguments on HN about how this could be good for the workers, thus successfully obfuscating the goal of screw-the-workers.
Although as far as that graph reflects the the study’s results, the new distribution looks almost perfectly balanced: even more people experienced a “10%+” bump than a 10%+ reduction post-update.
By what mechanism do you suggest the worker-screwing is happening here?
I'm not suggesting it is happening, the article is.
I'm pointing out that whether or not the mgt is trying to screw the workers, the opaque approach with the "Communications Team" generated only suspicion about both what was happening and the intent of management.
Let's assume you are correct and the entire adjustment was pay neutral - the total payout to workers for an identical set of deliveries was identical to the penny, only redistributed favoring/disfavoring different mixes of cargo and mileage. Why is it to anyone's advantage to hide that fact?
In fact, if you are trying to incentivize different behaviors, the best thing to do is to provide ALL the information on the reward structure, so the drivers can immediately read and analyze it and immediately adjust their selections to implement the new system.
Instead, the only thing management generated was confusion, ,mistrust, and poor implementation of their goals. For me, this raises a legitimate question of whether management is simply incompetent, or if they are trying to hide something (i.e., they're taking money off the worker's table and trying to avoid telling them).
even if total compensation was decreasing, the results for the company can be improved by being able to provide their services at a lower price point by cutting costs.
cutting costs is not "screwing workers". cutting costs is key to acting in a competitive market.
Yes, generically, cutting costs in the abstract is not screwing workers.
But when you got workers to sign up to work for you under Deal-A, and then you start reducing the workers' pay without their consent, you are screwing the workers. Especially so if you are surreptitiously reducing pay, so they cannot tell they are getting paid less until it is too late to choose to provide the work. It is simply dishonest.
Honest dealing would involve telling the workers up front, "we must cut costs, your pay will be x.y% less starting in two weeks; let us know if you'll be continuing under Deal-B". It would be the same if it was still "We're making total compensation the same as Deal-A, but adjusting it to reward tasks XYZ better and tasks PDQ this much less...". Honest dealing is saying it up front. Hiding the changes is less than honest and creates suspicion.
> I don't want to live in a world which is controlled unilaterally, and intransparently by a group of people who assume they have a full picture of the situation and assume they understand moral completely
I have thought about this topic for a while at the time that I worked with Law data (e.g. Family Law and Military Law), and I just came to the conclusion that several societal institutions and it's agents are inherently intransparent, even in situations where some "illusionist transparency" (there's transparency, but the magician deviates your attention to another side) is given (e.g. judiciary system, under-the-table political agreements, etc.).
That's one of the reasons I would like to have a more algorithmic society with human in the loop calling the final shots and placing the rationale on top. An algorithm will have human and institutional biases but in some sort, you can explain part of it and fine-tune it; a human making the final call on top of a given option would need to explain and rationally explain its decision. At best a human actor will use logic and make the right call, at worst it will transparently expose the biases of the individual.
<< That's one of the reasons I would like to have a more algorithmic society with human in the loop calling the final shots and placing the rationale on top. An algorithm will have human and institutional biases but in some sort, you can explain part of it and fine-tune it; a human making the final call on top of a given option would need to explain and rationally explain its decision. At best a human actor will use logic and make the right call, at worst it will transparently expose the biases of the individual.
I will admit that it is an interesting idea. I am not sure it would work well as a lot of the power ( and pressure to adjust as needed ) suddenly would move to the fine-tuning portion of the process to ensure human at the top can approve 'right' decisions. I am going to get my coffee now.
> That's one of the reasons I would like to have a more algorithmic society with human in the loop calling the final shots and placing the rationale on top
But isn’t that what the rule of law is supposed to be? A set of written rules with judges at the top to interpret or moderate them when all else fails.
The problem is that, for a variety of complex reasons, the rules are not applied evenly, and sometimes only enforced opportunistically.
So I don’t see how an algorithmic society is any different from today’s society. The problem is not the ability to operate algorithmically, which we already have, but in determining what the rules should be, how they should be enforced, what the penalties should be, who pays for what, and, perhaps most importantly, how to avoid capture of the algorithmic process by special interests.
None of these problems go away with an algorithmic approach, less so if there is a judge sitting on top who can make adjustments.
> a human making the final call on top of a given option would need to explain and rationally explain its decision.
To who? What you describe does not seem much different than the representation governments most of us here are accustomed to, other than the algorithm eases some day-to-day work required of the constituents. Already nobody cares, and no doubt would care even less if they could let an algorithm let them be even less involved.
>I don't want to live in a world which is controlled unilaterally, and intransparently by a group of people who assume they have a full picture of the situation and assume they understand moral completely, and also don't think it necessary to explain how they think so highly of themselves.
We already do; uncertainty is fundamental at all levels.
I think that's a very fair point, but wouldn't that be true even if Shipt hadn't made any changes?
It feels to me like the problem wasn't the change. For all we know, the change was a net good thing. The bad thing was the context in which the change occurred.
They had been transparent and it led to workers finding loopholes to make it unfair on each other. So there is value in opaqueness in that it's harder to exploit. But really a properly fair and transparent system would be better still.
That statement is wrong either way. Looking at the graph, ~22% got a >10% cut while ~36% got a >10% raise. Overall ~43% got a cut while ~57% got a raise. And if there’s any doubt, later on they dropped the “at least 10% less” qualifier for the 40% figure in text:
> But we felt that it was important to shine a light on those 40 percent of workers who had gotten an unannounced pay cut through a black box transition.
Can’t believe they (accidentally? intentionally?) screwed up the very first concrete figure given in the article. Guess what, discussion is now based on the wrong figure.
Also, the y-axis of the plot is labeled “number of workers” when it should be “percentage of participating workers”, unless they had exactly 100 participants (they say they had 200+). Lousy presentation.
I’m all for transparency, obviously.
Edit: In addition, since participation is entirely voluntary, common sense tells me that the data they gather should skew more negative, since people negatively affected are much more likely to participate.
it seems clear to me from first principles and from experience [0] that the result from voluntary compensation sharing skews low compared to the actual population. it's the people who want to complain about low wages who participate in such schemes.
that's not to say sharing wages is bad. this can still provide upward pressure on said wages. but the people who participate in the beginning are likely from the lower end of the distribution.
[0] websites like levels.fyi seem to consistently skew low when estimating higher percentiles
To clarify, I was basing the "some 30%" on the wage change distribution histogram that comes somewhat further down in the article from the statement you quote.
> Mostly gray area because gig corps don't like actually treating the workers as contractors, they just like the lowered costs.
I don't follow this? Is this predicated on the fact that gig corps can choose not to work with contractors that don't meet their criteria? If so, how is this different from only using lumber yards that consistently meet your expectations?
The implication of the parent poster seems to be that there are legal requirements regarding contractors and legal requirements regarding employees but gig corps would prefer to treat their workers as one or the other class depending on which is to their benefit - which would be against the law because of the aforementioned concept "legal requirements".
Exactly this - there are differences in what you can require from someone on employment contract and external contracting company (whether that company is single person or not) and effectively making one category into another without actually reclassifying (like employing as employee) is considered fraud in most places.
There are possible frauds from both employer and employee side, but I will list some common "landmines" of miss-classification, though beware that they are picked across different jurisdictions and I do not remember which apply where. All examples are possible items that can be decided to be part of misclassification, usually from contractor to effective employee:
- requiring specific dress code is non-enforceable on contractors in many places
- contractor is not required to provide specific person to fulfill the job, only a person of appropriate qualifications (it's valid for there to be a check on those qualifications)
- in UK case, contractor might be asked to prove that they have a substitute to work in their place!
- [Poland, possibly other] having only one client is not illegal, but can be grounds for investigations and if it's your only client where you work for equivalent of full-time job, it will be evidence for tax fraud
- You can not enforce working hours on contractors in most jurisdiction, only specific deliverables (taking part of work meetings is deliverable, requiring availability in general of specific person at specific times can be grounds for reclassification)
- above is often linked with "gig economy" - rules regarding "contractors" needing to pick up available jobs etc. are often considered illegal skirting of employment law.
As sibling comment mentioned, more is available from your local (too) friendly search engine. And employment lawyers and HR specialists.
In this case would it matter if they were contractors or not?
There is a way of calculating the pay for a job. It is predictable. Publish the algorithm. Want to change it? Great. Update the documentation and then publish that. The workers should be able to calculate exactly what they are owed. They can decide to leave or stay.
Only in America are people deflecting by bringing up the employment status of people when the issue is a lack of transparency designed to allow wage theft.
You’re arguing that it’s sometimes legitimate to hide payment information. So I can hire someone for a job, and only I know what the job is worth. The worker just has to try it out and see if their valuation of the job is the same as my valuation?
The fact that some, maybe most, are making more money is irrelevant. The organisation can change the value of a particular job at any time and simply say “the algorithm made me do it”.
But the worker knows exactly what the job is worth at the point of acceptance. Did you actually read the article? The issue is that the offered amount is decided algorithmically through an opaque process not that the worker doesn't know how much they will get paid for it.
A specific user can decide whether or not to take a specific delivery with full information. Following your analogy, it'd be like using some opaque algorithm to set the offer amount when hiring for a job, which is pretty much what happens today.
I don't think the reasons you are listing are valid at all.
1. If they can game it, they should. If that's a problem, then Shipt should fix their algorithm.
2. It's fine to data wrangle your way to an automated model for e.g. your company's growth projection or for predicting where you can best expand to find more customers. It's not okay to use it to unilaterally change pay agreements with your workers or "contractors".
3. The right of an individual to know what they're getting paid for their work outweighs any company's nebulous claim of its algorithm being a trade secret or something vague like that. Can you imagine businnesses actually operating like this? "You can't know what we'll pay you for this job, its ~~seeeecret~~. Just trust us."
It's amazing that anyone would even try to say that.
From now on I'm going to pay my rent according to a secret algorithm of my own devising and my landlord will just have to hope I'm generous this month. I can't have him gaming the lease agreement by only providing the things it says!
Why is this the case? It's well known that ML models are almost always gameable if you know the weights. What right do you have to their algo?
Why not? The contractors know before directly accepting a contract it's value, they can choose to not accept it if they think they're being underpaid?
Except this is how things work at basically all companies? Compensation decisions are secret and only known at offer time, I don't think I've worked for a single company where the specific executive decision reasoning for an offer is given.
Not one of those reasons constitutes a valid excuse. Holy cow.
"Surely in a hotly competitive market" there is no need to hypothesise about what "surely" would happen according to some wishful thinking. We know from countless past observation exactly what happens in any market of low skill low investment labor, wages go straight to the level of the most desperate willing to be slaves and make nothing at all, because there always are enough of those for everyone else to take advantage of, and having no money they have no power to demand better. They can't afford lawyers and ad campaigns and lobbyists and politicians, and they can't afford to strike, and that same wonderful magic market means there is no one better to work for, all employers are essentially the same. Unions and strikes do exist, some places where they are allowed to, and it's always a big news story and a miracle when they actually accomplish even 10% of what they needed to once in a blue moon.
This is straight up an unfair characterization. It's not "slaves making nothing at all", its fair value for their labor as decided by supply and demand. The issue you're highlighting shouldn't be solved by trying to legislate the market to oblivion which always introduces side effects. Just hand everyone a UBI payment a month and be done with it.
I didn't go far enough. Wages don't go right to the poverty line, they go right past it.
Because there are not only always people willing to work for net-nothing, there are actually enough people willing to work for net negative, because they manage to survive in various ways ranging from charity to theft. Mooching off someone else to live without rent or having to buy food, under the table jobs, minor crime, obviously health insurance is completely out of the question. As long as those exist, the market says that this work is worth this cost. The market doesn't know or care how that happens.
Unfair is a funny word to try to apply to the power-wielding benefit-collecting side of this dynamic.
Gig workers fall into a grey area between contractor and employee, depending on the particular jurisdiction and particular job. There are many ongoing legal battles trying to update outdated legal frameworks that don't quite account for modern realities. It is not as simple as you buying lumber.
Maybe not cut in stone, but if the data are derived from the workers’ payout receipts, it would seem likely that this was the amount that was paid out by Shipt. Or do you mean that third parties might be skimming something off after the company pays out?
I mean, it wasn't "a clear case of wage theft" as these weren't wage workers; and in any case it's not like they weren't told ahead of time how much the order would pay.
Do you know that for a fact they were told the total amount including tips, as well as all the details of the job, like items needed and delivery locations ahead of time? Gig apps often withhold info until after a worker accepts, to reduce workers ability to focus on desirable orders. In some jurisdictions Uber drivers don't see the final destination until after they accept a ride.
> For the shoppers who were being paid under the new algorithm, we found that 40 percent of workers were earning more than 10 percent less than they would have under the old algorithm. What’s more, looking at data from all geographic regions, we found that about one-third of workers were earning less than their state’s minimum wage.
> It wasn’t a clear case of wage theft, because 60 percent of workers were making about the same or slightly more under the new scheme. But we felt that it was important to shine a light on those 40 percent of workers who had gotten an unannounced pay cut through a black box transition.
Those paragraphs are then literally followed by a graph showing the conservation of fees overall but the redistribution of wages to other workers.
Seems to me that workers started cherry picking assignments leaving a bulk of orders which could not be realised, so they decided to (1) change the algorithm to avoid this, and (2) not publish it so it couldn't be gamed as easily.
