Uber, Data Darwinism and the future of work

81 thoughts on “Uber, Data Darwinism and the future of work”

  1. Very thoughtful article. Can you imagine what happens if low ratings tend to happen along ethnic, geographic or religious lines? How do our personal feelings and biases affect the reputation points we issue? What happens when we’re having a bad day? And how do we give people a chance to redeem themselves avoiding a truly vicious “rich get richer” situation?

    1. ST

      I think those are some issues we are going to confront but I be lying to say that I have any answer to your questions. Let’s put our collective thinking hats together and see how the situation evolves and changes.

  2. As you say, ratings and rankings will continue to influence more and more aspects of our daily interactions. Today it is very common to rank hotels and restaurants. In the hospitality business, we have experienced first hand the ‘nakedness’ of internet ratings. One employee (or one customer) having a bad day can cause potential customers to choose another establishment. There is no room for bad days. I heard it explained this way: ‘If you are naked, you better be buff’.

    experience first hand how No one feel bad about the effect this has

  3. i think you’re exactly right with this comment:

    > the challenges of the connected future are less technical and more
    > legislative, political and philsophical

    bitcoin is gaining strength. so is online reputation management. this could literally transform the nation state – we’ll be able to work from anywhere, choosing the political system we want to live under, while doing business in other jurisdictions.

    countries undergo a lot of political social engineering to achieve their goals, but they don’t always make for high quality living for those influenced by the policies.

    http://en.wikipedia.org/wiki/Social_engineering_(political_science)

  4. Suppose a few Uber drivers start to clamor for unionization. Will those drivers be cut out on the basis of their few “bad reviews?” Seems to me there needs to be a healthy dose of transparency in order to use data as the basis for “quantified society,” and not as the justification for unilateral application of a particular labor practice.

    I’m not suggesting that is what is happening here, but it’s very much worth considering the possibility for abuse and how it can be prevented.

      1. On the contrary, it is difficult to be narcissistic when you cannot meet rent or pay the bills. More so if you are homeless.
        Being tight financially teaches you humility.I am of course speaking from personal experience.

      2. Let us see how Cyprus plays out.

        The solution to the Cyprus bailout reveals narcissistic behaviour on the part of the politicians who want to ensure they are elected once again.

        As for the general population, not so sure.

  5. In a recent meeting at our office, the data cruncher told us what she found when she run a report. The manger run with it and declared an AHA moment. It was partly true because 70 to 80 % of the data entered was quantifiable. The rest, 20 to 30 %, we relied on observation of the client and ranking of the client by the account / client manager. What about race, ethnicity, language, religion etc…affect the observer and the observed? We have yet to transcend the dualities of life and then, maybe then, we can be unbiased. I changed so much, positively, in my opinion, observation and decision making process because of an experience I had during a few years stay in an islamic country and that is reflected now in my daily life including my work ethics. How does a program, a software or a tech gadget can measure and quantify that? Yes, I work in a non-profit sector where we, as a society, give people a second chance in social intercourse with the society at large?

  6. Reputation can be broadly applied to the class of social feedback systems. All such systems are subject to attack of the infosec kind. For the small price of N (target) + M (decoy) transactions, one can systematically inject malicious ratings against a target. That’s before looking an unintended bias introduced by the rating system itself, as any statistician will be happy to explain.

    The issue at hand is exactly why we have laws. Laundering opinion via an algorithm does not make it less of an opinion. If opinions were sufficient, we would be confident that suspicion is equivalent to sentencing to death, whether by robot or by denial of cloud-metered services, currency or labor dispatch. Since opinions are not sufficient, we will see new laboe laws which draw upon data as evidence. Which means the security/integrity of this evidence will matter enormously.

    A more pragmatic approach is to assume that the great reputation system in the sky is a benevolent dictator, or if you will, a parent. Why does society/law punish juveniles differently from adults? Because we allocate a grace period for learning, with an investment in measurements, social feedback systems, some records that persist and some that don’t.

    Any reputation system is a learning system. As long as humans are the ones learning, we must look to cybernetics for theoretical foundations, not to brute force machine learning, big data or artificial intelligence. Cybernetics has much to say about second-order feedback. The alternative is to surround humans with a tightening straitjacket of self-fulfilling Heisenberg-limited observations, until something pops.

      1. Om, Its ok for the whole idea to fall apart. There’s really nothing there. Adding dislikes doesn’t make it any more real. Just more buzz. Better some reputation system that showed # of visitors – and only # who said nice things; or # who would go back. You’d know all you’d need to know by seeing 10,000 people visited and only 6 had good things to say.

  7. I have stopped ranking except to say nice things. Why play god with someone’s livelihood? You don’t like it? Don’t use it again. Keep your opinions to yourself. Ain’t that a golden rule? “If you don’t have something nice to say, don’t say nothin’ at all.”

  8. In the absence of common value, ethics, and experience, we are forced to play God and that, is making part of our body, mainly the mind, bigger than the rest which becomes an ego. The schools don’t teach, as they use to and should, the whole body but only the mind which gets bigger and becomes an ego. Isn’t democracy quantified society? We saw the full spectrum of instant connectedness during the 2012 us election with all sorts of surveys and polls swinging the votes in all directions. Alas, let us be one with each other and technology, the next generation.

  9. This doesn’t mention the deeper thought: What happens when a very large sector of people posting online review are just assholes? Airbnb addresses this problem by letting both parties review one another, and that would seem to be a good model for Uber to follow. How fun would it be to see drivers review riders?

  10. Enjoyed this article very much for its valid points of how we are engaging (connecting) to advance better or more efficient services, but while endangering the personal reputation and eventual job opportunities for some. I have never used these type of services in the article, however can recognize how this social data phenomenon can become a device for prejudice. Likewise, are there rankings that rate the reputation of the person rating the service(s) used? For instance, reading through many ratings on Amazon, one can conclude some of the ow customer ratings are actually paid competitors, some purchasers who misuse products causing malfunction, etc., The best ratings have thorough reviews listing pros/cons and have details of how the product met or didn’t meet their needs as advertised. Maybe instead of just tapping buttons for rating a service, a voice recording will be used to authenticate a valid rating instead of having anonymous screen names or customer id #s.

  11. Reputation online is definitely only going to get more important in the near future.

    It’s interesting to think about how it will be standardized across platforms as well. That’s where these issues can become amplified.

    If you’re a good Airbnb host, should that not translate to trust on TaskRabbit? If you screw someone over on one platform, is that not relevant for another collaborative consumption platform?

    Reputation is also separate from trust. Just because students didn’t like your class on Udemy, doesn’t mean you aren’t trustworthy as a person.

    I think someone will develop a very compelling solution to standardizing reputation online. I know a number who have been working on it but no one has come close to figuring it out.

    And when it exists, the way we all work and interact with each other will never be the same.

    -David
    (Worked on Community, Trust and Safety at Zaarly)

  12. Great post and I couldn’t agree more that a “state of connectedness” is growing but that the future of it won’t be about enablement. It will be about the ethics, philosophy, and legalities of doing so. It’s the reason I revamped my own blog to [RE]Think (www.rethinkeverythingblog.com) and am working on tackling the issue (amongst others) of defining “human” in such a state of connectedness. You broach this in the end of your piece, begging the ultimate question, “are we still human?” when we are always connected? Are we less than human if we aren’t? What does it mean to be human when we have both a digital self and a physical self? Do we need to redefine what it means to be human (both for ourselves and for our species; this is the position I’m taking in the book I have just about finished)? I love the idea that you are digging into this (even if from a more practical angle) because it will take people like you to tackle these thorny issues. Hope you write some more on this topic soon.

  13. On a tangent, I see a new kind of union evolving where micro-entrepreneurs would use data to negotiate / bargain and may collectively carry a ‘kill-switch’ to lock such a network out, akin to how traditional unions do a physical lockout.

  14. Yes.

    It’s not just that yellow tax drivers who paid for medallions and rent medallions, or limousine drivers, get screwed by Uber’s pretend “market competition” (achieved by practices like this).

    It’s that it’s one more brick in the wall of all of us becoming human Mechanical Turks and Internet subsistence farmers supporting the peer-to-peer hucksters:

    http://3dblogger.typepad.com/wired_state/2013/02/internet-subsistence-farming.html

    This is what I said after the enthusiastic coverage by Forbes of Air BnB, that same lovely “disruptive” force:

    They are saying: Oligarchy for me, technocommunism for thee

  15. This is really good. There’s a lot of talk of reputation, quantification, transparency. I’ve written about these before but I like where this is going. Do you think apart from a “Quantifited Society” we’ll also see a “Quantified Enterprise” ?
    Something built on all of those concepts but both consumer and employee has access to that transparency and visible reputation ?

    Would it mean an organization is only as good as the sum of its collective reputation ?
    Would it mean that consumers could very well choose who they deal with on the inside based on individual employee quantification ?
    And vice-versa, will companies choose just who they want as a customer based on transparency ?

    It could go either way, and each has its own impact.

  16. Related: Mechanical Turk made a drastic change recently – it no longer accepts workers from outside the US. Why? Because of widespread quality of work issues combined with challenges in tracking individual performance.

  17. It is apparent that some on-line activity is sourced from PR sock-puppets and lobbyist astro-turfers (as well as Government-operated spoofing – eg US Air Force persona-management RFP).

    If a “reputation system” gains the influence/authority to cause individuals to lose their employment, it seems crucial to ENSURE that the feedback is authentic: originating from genuine customers.

    Without such safeguards, one could leverage the reputation-system to target enemies. For example, it would be easy for management to remove uppity employees (such as union organizers) simply by hiring a covert sock-puppet operator to generate negative feedback. The potential for wrongdoing by government sock-puppets is equally great.

  18. Very well done Om. Recently had an acquaintance lose her job as a bartender after a fairly pointed Yelp review and it struck me that your post has shone light on this dark (not so cool side) of the digital pilliow. A place where due process subjugates itself at the alter of immediacy and transparency.

  19. In terms of Yelp and the generation we’re in now when tons of businesses and people are “rateable” online, will we ever reach a time when these ratings won’t be taken with a grain of salt?

    For example, if I read Online reviews of a restaurant, I must keep in mind that these people are not like myself, in that I never write reviews for Yelp. Add to the fact that they’re largely anonymous, how should i take this info? And who’s to stop competitors from anonymously bashing each other online?

    I think overall this review-everything generation will force everyone to take no easy days in the office – esp when dealing with customers. Everyone becomes extremely accountable for every action. Very capitalist and Darwinian indeed. Great piece, Mr. Malik.

  20. As dotpeople pointed out, something to be cautious of would be cyber bullying. For example, in the above system, you could threaten to post a negative review and have all your friends post negative reviews if someone doesn’t do what you want. There should be a way to verify a negative review by matching it up with a service so it would be validated. This could provide a measure of protection for the employee, from unjust dismissal, and the employer from losing a valuable employee because would could essentially be considered malicious gossip.

  21. Great article and kind of have potential to trigger debates or preparing us for quantitative society, which in a way might be good, as can create more transparent society that what we have now but agree sometimes numbers misses context and so far we dont know how we can merge both numbers and context until we do some wonders in semantic analysis! http://dearsrk.wordpress.com/

  22. This quantified metric plays a really big role now days in the service industry. My cousin works at a restaurant and during their morning role call they go over Yelp reviews. If there is a negative review the management scolds the team and tells them they need to step it up. If you get specifically called out in the review for doing something wrong you could be fired. It happened with a few employees already.

  23. Om, this is EXACTLY why I choose to follow your blog way more than many others. It is also why I comment here often. Not only is your discourse spot on with implications, but I think it will establish a connection between the era of ‘good enough,’ that is ending and the start of the reputation, ‘be the best,’ era to come.

    it will be based just as much on interaction as it will be quality of service and as you say those that keep that detail in mind will do well.

  24. We already have a system that has many years of successful reputation management – ebay. It seems to be a fairly accurate system and is bi-directional.

    As posted earlier, there should be a mechanism to validate that the evaluations are based on a legitimate transaction. Perhaps Yelp could put a key code on every receipt, for example.

    As to the larger problem of the growing influence of public or online reputations, society will adapt pretty quickly. The challenge is the low cost of accessing and ubiquity of the ratings — it discourages spontaneous assessments. The public benefits in better service levels, lower economic friction, etc., however, will outweigh the negatives.

  25. There’s this assumption, which I think is unwarranted, that the ratings will be real and not submitted by spammers, competitors etc.

    Or even with people with radically different likes from you, e.g., yelp (I’m well north of the median age of their reviewers).

    Getting rating systems that convey what “someone like you” thinks is the key.

  26. I wonder if the main problem here is putting too much faith in not enough data? I think a feedback-driven work environment makes a lot of sense, but a single “how did you like your trip?” rating lacks a LOT of context, and creates possibilities such as the current Uber situation.

    As many people above stated, bi-directional ratings would be a great start. Simply being able to filter ratings by the score of the person giving them would help insulate workers from unreliable low scores.

    Another thing that’s worth thinking about, what other metrics could/should enter into the equation? What about accident history? Or traffic violations? What about data around continuing education based on the feedback? For instance, what if a driver was new to an area and received low rankings because he got lost often, but then was taking a course to better familiarize himself with the area, and somehow that data was available when making decisions about who to prioritize in the system? Would that be a meaningful addition to the algorithm? I think it would.

  27. Om – fantastic analysis of a still nascent world. Personalized analytics are the future. There is a lot to work out. I’m the co-founder of a reputation & personalized analytics company (Virtrue.us) – we see and debate this issue a lot. Our conclusion, as things stand today, is that you are your data. In fact, the online view of you is probably more accurate than the one you may want to present to the “real world” public. What this ends up meaning is that you have to be careful with your data and be careful how you interact online. We are trained to be careful offline – i.e. don’t commit a crime or else the police may catch you or don’t go into a certain bar because people might associate you with the a certain group. This care forces us to act in certain ways (laws having their mostly intended affect). These online laws do not exist so people feel free to act how they want. This is changing as the Uber situation highlights. Acting one way (even in a “real world” setting, albeit a quantified one) when you thought you could hide has real consequences. I believe that most govt laws will be slow to catch up (if ever) and, instead, data darwinism will do exactly what the name implies.

  28. Great, thoughtful article. I see this all the time on Twitter, Yelp, Taskrabbit, and other services: affluent, upper-middle class (frequently of a certain skin tone) professionals compalining about bad service and miscommunication, running the risk of seriously ruining someone’s livelihood. It’s not worth it to me. Unless service was good or seriously, seriously bad, I just don’t rank. Period.

  29. Great article and interesting point. I would argue in the case of Uber we’re also seeing changing user behavior. Across the service industry we’ve always had a very efficient feedback mechanism – tips, but management had no (real) visibility into the data. Now with Uber, they’ve implemented have a different feedback mechanism and have taken tips out of the equation. I don’t have the option to tip, so now I rate – 5 stars on Uber is the new 20% tip (without the option to go higher).

    Personally, I think this is the worst part about Uber (otherwise I’m a hug fan). They’ve removed tipping which drives quality service. Now, if I have a good or bad experience with a driver I don’t get to adjust the tip, so I adjust the rating.

    I’m sure drivers love this because they’re guaranteed a healthy tip, of course the ones that got let go for bad ratings are upset. But this is a bad tip at scale. I don’t know what the process is for informing the drivers – but it seems like they get plenty of visibility into their ratings (http://thenextweb.com/insider/2012/04/13/in-case-you-didnt-know-uber-drivers-see-how-many-stars-you-gave-them/)

    I wonder how this will impact other service industries as our current societal norms like tipping are changed by technology. Will a public rating be the only way to reward or punish behavior?

    (Also, a couple other commenters made statements that Uber should allow drivers to rate passengers, my information may be dated but I was under the impression this is built into the system).

  30. Interesting comment dotpeople made. It was made a little too high-mindedly, I think; it almost seems designed to say “look how smart I am for using fancy words and academic concepts”. So let me re-state the two salient points he/she made:

    1. The “Yelp Problem”. It is far too easy to launch a scathing “artings attack” for reasons that have nothing to do with actual performance.

    2. The “blunt instrument” problem. We still have not really progressed beyond 5-stars. This is certainly understandable. Collecting nuanced feedback is much more challenging in general, and especially when you rely on volunteers (i.e., customers) to do the rating. In our online world, these volunteers are not only unpaid, they are usually PAYING.

    3. And from the latter, “response bias”. I’d suggest that many if not most people who would give that 3-star (middling) rating will just abstain, leaving a very skewed result.

    We’ve really screwed ourselves with this, I think. I don’t see how it could have been avoided, but I find myself wishing we could stuff this genie back in the bottle. And I am a tech entrepreneur!

    1. Thanks for the feedback, no design was needed 🙂 Typed the comment and pressed send before I could talk myself out of posting. Appreciate the translation.

      > We’ve really screwed ourselves with this, I think. I don’t see how it could have been avoided, but I find myself wishing we could stuff this genie back in the bottle. And I am a tech entrepreneur!

      It is still early, and the systems are redeemable. How do we “escalate” co-worker performance issues in a large enterprise? Usually via some form of hierarchy, official or unofficial. Yes, those hierarchies create endless corporate politics, but they also provide time/latency for learning and performance improvement.

      The new “flat” reputation systems are deservedly excited about the hierarchies they are replacing, but new hierarchies are inevitable, as George Orwell’s Animal Farm timelessly illustrates. The good news is that it will be much easier to “reorg” a database-driven reputation hierarchy than to reorg the IBMs, HPs and Dells of the world.

      For consumers, there can be hierarchies of reputation systems, along with tools to “compose” reputations across services. The technology (interoperable annotation servers, http://www.openannotation.org/) will soon be here, but it will take a while for business models to catch up. See also http://hypothes.is

  31. Om, great piece that really gets me thinking about what connectedness and so much data bring about. The biggest changes coming in the next few years aren’t the technical ones…they are the societal, legal and cultural ones. The ramifications of reputation that is often controlled and sometimes beyond our control is deeply personal and as you showed, potentially fragile. If reputation is the new currency, how far will people go to create, protect or destroy it?

    We wrote up our thoughts here:
    http://successfulworkplace.com/2013/03/18/data-and-connectedness-create-a-reputation-wild-west/

  32. The ESSENTIAL point to keep in mind is what the rating means to all involved e.g. The Rater, The Rated, The Interpreter/Enforcer. How will you feel when you realize you just FIRED that hard working but not very handsome, polished, Brooks Brothers dressed driver with your 4 star rating? You thought he was a GOOD driver, not a perfect one. You do NOT feel he has done anything to deserve being prevented to work to feed his children. Tough Luck! You just fired him. Can you sleep at night?

    1. As a society, we can draw new lines on information flow.

      Some have claimed that data liquidity implies an inevitable loss of privacy. If that is true, why aren’t Apple’s iPhone5 blueprints available via web search? Because their economic value justifies the cost of protection/privacy/security. Not to mention the consequences that have befallen prospective violators of Apple’s corporate firewall.

      Imagine if the list of people who were fired as a result of your 4-star rating followed you around everywhere? In the service industry, on social networks, in employment reviews, in the elite school applications for your children? Feedback can flow in all directions. Is that good? Maybe.

      We are like a species who have discovered the freedom of flight for the first time. We initially measure, store, search, blacklist, whitelist and analyze to our hearts content. Eventually we’ll step back and ask ourselves to formalize the social objectives of our algos.

  33. I am having a hard time making a decent argument on behalf of UBER firing a driver on the basis of a poor rating of 4.5 * when it’s own rating on YELP is below the driver’s at 4.0*. Am I missing something? Will somebody please help explain this?

  34. The myth is that given sufficient numbers a ranking will be objective: the wisdom of the crowd.

    This is mythical because it is exceedingly rare that the numbers reach a level to make them objective. Look at rankings of hotels or books or anything else. Numbers almost never exceed 100. I worked for a company with 15,000 employees which exhorted us every year to vote for the company’s product in online rankings. How hard is it to mobilise (or fake) 100 voters?

  35. ammy Pena “Connected began in 1996 making those who depend on a sense of complete “connectedness” under the age of twenty or old enough to take responsibility for stock market crashes, bank bail outs, and sinking a perfectly good Twinkie in the year 2012. I would gather to say that with wisdom we choose to be less “connected. Will this lead to a complete social injustice where people are unable to identify themselves with anything????

  36. So much for copy/past from my smart phone. I failed to include the letter T in my first name and the teeny tiny end quote after “connected”. The statement is real and the flaws can be easily corrected – T

  37. This post reminds me of the movie Gattaca. Maybe that scenario would be better than a data-driven class system. At least our species would be pushed towards being healthier, which in the end is the only thing that matters.

    Thanks for giving me something to ponder as I go to bed tonight. Some day, I’d like to CentUp brilliant writing like this.

  38. The worker is the new entrepreneur ,two way universal ratings and rankings for both the employed and The employer..
    you own U – Freedom to choose, charge, manage , scale & operate, creating value for your work, building reputation and earning respect.
    Not a fantasy or rhetoric a system coming into effect soon..

  39. Great article…some quick thoughts, hopefully not entirely redundant:

    I agree that the problem with “likes” and vanity metrics like “star ratings” that are not tied to verified transactions or identities, is that they are both subjective (my 4 stars could be equal to your 3 stars, or go up and down based on mood), and too easily manipulated (I’ll give you a discount if you “like” my page or rate me 5 stars, or I can falsify/spoof ratings to move myself up and a competitor down.)

    At TaskIT, we spend a lot of time thinking about ways to leverage these metrics and the growing spectrum of subjective social data that factors into a buying decision, while avoiding or mitigating some of the issues. For example, we look for ways to augment the subjective data with objective (verified, quantifiable) data and metrics from our own transactions and via trusted 3rd parties, and leverage true identify, relationships, endorsements, and other socially authentic and relevant datasets that all assist the buyer in making an informed decision.

    The data is only as valuable as the trust it inherits. An anonymous dislike or low star rating from an obviously disgruntled party doesn’t mean very much, and should have negligible impact on the product or provider being rated. But as the data set grows, interesting trends start to emerge. Numbers do matter, even though they don’t guarantee objectivity. Percentages matter (ratio of bad reviews to good ones). Identities matter. Details matter. Authenticity matters (i.e. tying a review or rating to a verified transaction).

    These are not new issues, but the scale and impact has sure changed. A bad review or reference (whether formal or casual, online or offline) has always had the potential to cost you a job, a customer, a date, etc… But now that we are leveraging large datasets to make data-driven decision about all aspects of our lives, the data has to be interpreted with context, it has to be authenticated, verified, even scrutinized (i.e. open feedback on a bad review).

  40. As much as I long for idealism, the constant connective environment is reducing the friction and delay in bad or at least less than ideal behavior. “Life is like high school,” said Harry Beckwith in the book Selling the Invisible. Now, we seem to have the drama of high school culture, set to high gear with no pause.

    Constant connection seems to be denying us the chance to breath, ponder and reconsider. The raw nerves and instant reaction times are feeding into a sort of hyperactive unreality. So intense. So unforgiving. So judgemental. Our politics, our journalism and our business ethics are reflecting this growth of gracelessness. It will be interesting if that is the new normal.

  41. Very interesting article, and I love the term data darwinism. Just as we have to have media literacy, I can foresee a need for statistical literacy as well.

  42. This is an incredibly relevant topic. We at WorkMarket struggle with it everyday. We have a marketplace of hundreds of enterprise buyers of on-demand service providers and we have a database of 20,000+ freelancers, contractors, and companies that are getting ranked, rated, and analyzed for every single assignment sent. We are constantly in the inquiry of what is fair, how should ratings work, how do we provide transparency and analytics around reputation and reliability, etc? It is a constantly evolving conversation and I am in full agreement, that the quantitative measurement that technology enables will completely redefine the world and society that we are now seeing change before our eyes.

    1. @jamesd are there any lessons you can share about how you guys have handled the freelancers and the rankings etc. I would love to find a way to share those with the world at large. Please get in touch.

  43. Great stuff Om. Thanks. Question. You write: “In the industrial era, labor unrest came when the workers felt that the owners were profitting wrongfully from them. I wonder if in the connected age, we are going to see labor unrest when folks are unceremoniously dropped from the on-demand labor pool.” What’s the difference? Isn’t the latter exactly the same as the former?

    1. It is pretty much the same thing except for one small difference: the justifications are going to be different and there is going to be a lot more data-backed arguments. I am following up with another post in a few days after I have had some time to ruminate on things.

  44. One idea that seems out of use these days is the “Global Village” term – but it appears more and more in practice.

    Take a step back and imagine a village/town small enough so that everybody knows practically everybody else.

    Imagine the Uber situation there – if there are two cab drivers and people know that one gives much better service than the other; then the situation is simple – if there is enough demand for both of them, then they will not get equal rates; and if there is enough demand for one of them, the other will be permanently, irrevocably out of his job and will have to do something else.

    Imagine the supermarket or Amazon purchase data mining there – the merchant knows all about what you buy, what you like and will be able to offer detailed recommendations on new stuff. Also, the merchant knows if, for a very crude example, a local priest is buying gay pornography; and just as in the modern digital world, any limitations on what to do with this information are legal and ethical, but the capability is not restricted.

    Et cetera. The social interactions in the upcoming world IMHO really are much more related to a [global] village rather than the anonymous, hiding in the masses life in the pre-digital 20th century megacities. Initially the internet was even more anonymous than those cities (i.e. “on Internet, nobody knows you’re a dog” meme), but it’s not like that anymore and won’t be ever again.

    1. One difference is that small villages are illiquid markets. It isn’t as simple as “choose the best cab driver”. What happens at rush hour, when your choice in cab drivers is limited? Or the bad cab driver is your brother-in-law? Or the father of your child’s best friend? It’s non-trivial y to externalize undesirables to a neighboring village’s backyard.

      Information cannot be separated from power. Everyone may know (“information”) about the local priest, but if the priest also owns the bank, school and church, that information may have limited impact (ask Matt Taibbi). Few information systems are symmetric, because power/assets are not balanced. Market makers tussle with market players, needing each other for scale, but always testing power boundaries.

      When an information system overlays a physical city to aggregate liquidity for a virtual market, where should the rejected players go? Into an education market for remediation? To a lower-priced market that cannot afford high-performing players? In China Meiville’s “City and the City”, two virtual cities occupy one physical city, and natives are taught from birth to “unsee” each other. How many unseeing virtual cities can we overlay onto a single physical city?

      It is the scale of megacities and their supply chains that enables the arbitrage that funds quantified data collection and analysis: http://qz.com/64560/whos-to-blame-for-see-through-yoga-pants-and-horse-meatballs-the-independent-republic-of-the-supply-chain . At city scale, many of the small village’s self-stabilizing social mechanisms are absent. Can we reverse engineer and scale “village community” as fast as we scale blacklisting?

      Quantified feedback loops are a form of involuntary cooperation. The question remains: who (power) defines and revises the rules (information) of each system? We create systems and live in them, then we find ourselves changed by our systems. Should we record time capsules of our intentions at the time we entered the system, in case we later want to reboot and explore an alternate future?

      We may look at a bell curve of quantified performance and draw quality boundaries. We may scramble to negotiate our position on the good side of a boundary. Or like the Star Trek Kobayashi Maru, we may embark our quantified selves onto a creative quest for markets with less arbitrage, more symmetry and more human hope.

  45. First, any manager that would fire an employee for performance/ratings without fair warning is a poor manager. I see enough of them in corporate America, and the lack of consideration is sadly entrenched. The question we should ask, is what happens when we begin to also apply this to said managers and companies? The 21st century is going to be quite unstable in the measurement of value and potentially money, ie. Bitcoin.

    Aren’t companies already subject to this pressure? Is viewing the ratings of products before purchase already changing the shopping relationship? I would hazard to guess that failures in physical commerce is the result of the breakdown of selection and price against value judgements including reviews. Bigger stores used to be successful because they offered choice. Now they are dinosaurs because they offer choice without value.

  46. Wow, this really resonates with something I wrote in the last few days about the limits of networks to solve all our social problems:

    http://www.huffingtonpost.com/catherine-bracy/whats-progressive-about-peer-progressives_b_2861146.html

    It’s sort of in response to Steven Johnson’s Future Perfect. I wish I had known about this Uber issue when I was writing–I even posit the hypothesis of Uber drivers organizing against the company!

    Anyway, really important issue, something that needs to be taken into account when we think about what “tech policy” means.

  47. A customer service rep for some company I don’t remember, handled my call and then asked for company purposes how good the service was and I replied 4/5. This appeared to be an unsatisfactory number and I was asked why I did not give 5/5. Why should I give top rating when a person just does their job and did not do anything special or extraordinary?
    A while back, ATT phone services used to do a feedback survey asking how the customer service rep handled our problem. It must have been pretty bad because the reps would deal with our issue and then beg for a top rating, even give a specific instructions on what to do when the feedback phonecall would come the next day – “So you will have to press 1 when rating me.”
    I told my husband these people sound desperate for them to be begging for high ratings.

  48. You say you start the day connected and end the day connected, That resonates so much with me as what I like to think of myself as in Ireland ~ a Social Bridge.

    Look forward to reading more and so glad I’ve found you!

  49. Not addressed – but critical to the topic – I can tell you from years of personal experience in community organizing, trade union organizing that nothing is more important than individual dignity.

    If you are treated as unimportant, only a cog in the machine producing profits, a robot only required to obey and not contribute in self-conscious fashion, there cannot be loyalty or real satsfaction. Alienation hasn’t changed since the 19th Century. Ownership defines attitudes towards producers on the payroll – and still tends to come down on the side of arrogance and caprice.

  50. Reblogged this on Business Travel Technology – Today, Tomorrow and Beyond… and commented:
    This a great piece on the challenges of the connected on-demand workforce that exists today. As business travelers we constantly review and comment on the goods and services we consume. Never before have we been at a time when instantaneous commentary and review can be provided on things, services and even people. Something to think about when we click on the review/rate button in our browsers or mobile phones…

    ~Jim

  51. very thoughtful article, and I like the comments regarding narcissism in the 21st Century..as a writer of content I begin to judge my content by the number of comments or “shares” I get, and fall into the trap of trying to outshare the last article.. sometimes good content is just good content and you wonder if this data darwinism you speak of could just be a fad that the connected generation overcomes when it gets bored with its new toys…

  52. I found your article both timely and insightful. The questions you’ve raised are ones that I have been asking since I began working in crowdsourcing and speaking, writing, defending and lobbying on the labor issues it raises. I think there are ways to begun to construct social norms and rules that will support this way of working in a connected world. I’ve worked on proposed legislation to try to clarify some of the labor laws to better enable this type of work. the issues raised by Uber’s drivers are very typical of a disgruntled freelance workforce, and with the rise of data darwininsm, as you so eloquently have termed it, those of us who are in the industry and passionate about its success must do more to force these conversations. I would be happy to share my experience and thoughts on how to help society move through this intricate web of law, policy and social constructs.

  53. How is user feedback via highly connected technology any different than our notions of capitalism? I find your argument (in essence that connectedness/user feedback weeds out bad products/services) to be simplistic and binary. In an efficient capitalist society, the success of a product or service is determined by the crowd. Valued products and services survive, others go by the wayside. However, you need to fold in some oversight via government regulation. Certain products and services need to be more tightly scrutinized by a governmental body to ensure safety, consistency, availability. Transportation, for example, is important to more highly scrutinize. Without oversight, markets can more easily tend to imbalance (imbalances that cannot be well-managed by a company with only $49.5M in venture capital funding). Imbalances can lead to safety concerns (age of the taxi, age of the airplane, age of the bus). In general, Uber does a fantastic job setting customer expectations about the availability of a ride and Uber does a fantastic job of making payment simple. Blame is mitigated by Uber when it comes to discriminatory driver behavior, or unsafe driving, or unsafe vehicle, or assault on a passenger by putting that responsibility on the driver (that’s why oversight and standards are important). Uber is a technology that matches you up with drivers (and does so well for certain people, with certain devices, in certain parts of town, at a certain price point), but can Uber possibly do the job of the Public Utilities Commission, the SFMTA, NIST, FTC. Bottom line is that nobody says as their plane crashes to the ground: “At least I could track my plane on a map and the payment process was seamless……” That’s what the FAA is for.

  54. Seems like the etiquette is simple and timeless: Do all you say you will do. Don’t encroach on other’s or their property. Make amends quickly.

    With reputations, have people put their money where their mouth is and include a simple, local, expert and fast conflict resolution process.

    Seems like all these things are just resources. So manage them like you would any other resource.
    http://wp.me/p1ePZy-1ri

  55. I think in the very large scheme of things, ratings are to be expected. And there is no reason to panic. Because the system will find a way for self-correction. For example, raters will also build up a creditability based on their ability to rate judiciously.
    The key thing to remember is, collectively, this means we will become a better species.
    That is all that matters.

  56. I think it’s interesting that there have been very few comments thus far on the way that reputation works for employers and companies.

    When a company mis-uses data to fire workers on the pretext of poor ratings when the actual reason might be a labour dispute or something completely unrelated to performance, the company may start to get a bad reputation as well. And may have trouble hiring people. Or getting other companies/governments to work with them.

    I think there is also a laziness at work here and an over-reliance on raw data. When the only tool you work with is the hammer of raw reputation data, everything starts looking like a nail.

  57. Most companies pay to receive feedback from customers and then attempt to absorb that information in order to do something better. The idea that services, products and people are continuously evaluated and ranked may simply increase the velocity of evolution. (It reminds me of running one of those evolution simulations that Dawkins had in an early book). With enough data, a bad day may not be significant. So it might not be at all bad, particularly for individuals who are willing and capable of working hard. Providing ‘appropriate safeguards’ are put in pace. And this is the tricky bit. Someone once said to me that ” .. if people want it, and technology permits it, then the regulator will ultimately allow it … ” I think that the connectedness of us all will create inevitable changes everyone; consumers, students, lawmakers, ceos, businesses, workers, politicians & governments – it is an unstoppable force. The big question to me is how to define these appropriate safeguards and enact them swiftly enough in a rapidaly changing environment. Most of the smart people are outside of government – just try imposing bonus restrictions on bankers!

  58. I like how you mentioned that you think about your ratings, even when it’s for 3 stars. However, most responses/ratings are polarized. People tend to rate and comment when they have had either really awful or really great experiences. And the quality of the comments of 3-star ratings and their equivalent tend to be wildly different from the quality of the comments of polarized ratings. Relying on a biased sample of ratings to build a workforce or an economy would be…problematic. At the very least, we need to talk about sample sizes

This site uses Akismet to reduce spam. Learn how your comment data is processed.