Piecework, Political Economy and the Internet

by Henry Farrell on April 1, 2008

This “piece”:http://www.portfolio.com/views/blogs/market-movers/2008/04/01/blogonomics-valleywags-pay by Felix Salmon on the problems that Gawker Media is encountering with pay-per-pageview is pretty interesting.

Golson’s take-home pay is so much larger than his base salary that his base salary ($2,500 a month) has become basically irrelevant. Instead, he’s been relying entirely on his PVR of $9.75 per thousand pageviews – a rate which has seen him taking home more than $4,000 a month so far this year. For Golson, then, his realistic base salary is in the $4,000 range – much higher than the $2,500 which Robischon is referring to. … The problem here could have been partially fixed if Robischon had decided to give Golson a more realistic base salary to begin with. But Robischon’s boss, Nick Denton, wants fixed salaries to be as low as possible: he hates it when a writer doesn’t justify his salary with pageviews, and the best way of ensuring that situation never arises is to make the fixed salaries as low as possible.

This PVR is being lowered, leading to a strong reaction from Golson and others. Salmon explains their anger in terms of psychological mechanisms such as loss aversion, which are indeed applicable. But I think that there are two other things going on, both of which have to do with the economics of piecework. And after all, paying people on the basis of the number of pageviews their articles receive is a glorified version of piecework.

In many ways, piecework appears to be a more efficient means of pay than, say, providing a fixed salary. It would appear to align the money paid to the employee more closely with the revenues generated by that employee’s additional effort than would conventional wages. Hence, I presume, Nick Denton’s detestation of big fixed salaries – they don’t necessarily maximize page-views and advertising revenue. You might expect that in an ideal world, all employees would be paid on the basis of a particular form of piecework pay – that is, they would receive compensation directly commensurate to the revenue generated by each piece that they created.

But there are two problems with this. First, the revenue generated by each piece may be affected by factors extraneous to the effort provided by the worker. If the revenue generated were, for example, to increase for purely exogenous reasons, the compensation received by the employee would also increase, without any additional effort on his part. From the point of view of the employer, this would be a simple redistribution of profit from herself to her employee without any compensating benefits in terms of increased effort etc. A rational boss would seek to resist this.

The second problem is that the employer, even apart from these issues, has difficulties in committing to reward piecework properly. WUSTL political scientist Gary Miller’s _Managerial Dilemmas: The Political Economy of Hierarchy_, a book that I’ve repeatedly recommended, gives a nice account of this problem. If employees provide full additional effort at time _t_, providing a nice bonus for both employer and employee, they have, in effect, revealed the maximum degree of effort that they are able to put into the work. The employer then has an incentive to lower the piecework rate at time _t+1_ so as to increase her share of overall revenues, while demanding that the employee continue at the previous rate, or be fired. A rational worker will therefore not make full effort at time _t_, figuring out that he will be screwed over later if he does. Miller discusses case history evidence that suggests that this really does happen. This helps explain the fragility of piecework systems – while they are in principle more efficient, they only work properly if they are supported by interwoven set of commitments and expectations that are quite hard to maintain in a capital-dominated market economy. The political economy of hierarchy, as Miller says, is important to explaining why. This is also part of the reason why most of us get wages rather than piecework pay (another reason is that the value of our output is often difficult to measure directly, but that is another set of issues).

From a superficial reading, it would appear that Gawker Media faces both of these problems. First, the exogenous factors problem:

The Denton plan is particularly brutal in the case of Wonkette, where PVRs are being slashed quite dramatically this election year. On a logical level, this makes sense, but on a psychological level it can be extremely demoralizing.

Wonkette is obviously going to get more pageviews, as do most politics oriented sites, during an election year, without any additional effort from the employees. Denton here is acting to make sure that the windfall surplus goes to him rather than his employees.

Second, the commitment problems.

If things go according to Denton’s master plan, Gawker Media writers see their take-home pay rise steadily from month to month. Every three months, there’s a small step down, but over the long term the secular increase in pageviews more than makes up for the quarterly decrease in PVR. … It turns out that Golson got 557,469 pageviews in March, which equates to a total paycheck of $5,435. That’s well over double his base pay. His colleague Nicholas Carlson earned $9,025 for the month, in which Valleywag as a whole got just over 5 million pageviews. That’s an impressive rate of growth, but it does help explain why Denton might want to bring Valleywag’s PVR down towards the levels seen in the rest of Gawker Media.

Here, I don’t think one needs to invoke mechanisms of loss-aversion (although they may play an important role in practice). What appears to be happening is that the pageviews for writers working for some of the Gawker Media websites are rising substantially over time, even apart from the conjunctural factors affecting Wonkette etc. Presumably, a significant amount of this is due to the employees’ efforts; up to now, they have been benefitting substantially from those increases in pageviews. Denton is effectively trying to reset the compensation system so that (a) employees’ overall compensation only increases at a relatively slow rate, but (b) they are still incentivized to produce more pageviews, because most of their compensation will still depend on how many pageviews they produce, while, most importantly, (c ) they will get much less for each individual pageview than they used to under the old system, and Denton will correspondingly get more. This seems to me to be exactly analogous to the commitment problem that Miller describes. It’s entirely unsurprising to me that employees are up in arms about this, even apart from loss aversion, as it implies that any further productivity gains they make are likely to be pocketed by the proprietor rather than returned in part to them. If they had been rational, perhaps they wouldn’t have busted their asses to help produce this outcome in the first place, but in any event, I suspect (I could be wrong) that Denton is about to be introduced rather abruptly to the disincentive effects of broken implicit commitments.



mpowell 04.01.08 at 10:14 pm

Yeah, I sure know how I’d react. Unless Denton can pretty clearly demonstrate that $9.75/thousand page views exceeds the marginal income for that many page views, I’d be raising holy hell about this. Even if this is only related to a seasonal increase in page views, why would I be willing to accept a longterm reduction in my PVR?


dsquared 04.01.08 at 10:15 pm

I would call this the commission sales model, as under piecework you jut get paid for what you produce and the employer takes the marketing risk, but I suspect that the distinction here might not be sociologically important as the “rate busting” issue certainly happens under both systems.


Kieran Healy 04.01.08 at 10:22 pm

It’s striking how prevalent throwaway analysis in Kahneman/Tversky type terms has become (using ideas like loss aversion, anchor points, etc) even when these concepts don’t really have a strong connection to the problem at hand. While these are important mechanisms, reducing problems like this to questions of psychological interpretation seems mostly just to strip out questions of social context — in this case, the structure of employment relations.


dsquared 04.01.08 at 10:44 pm

If I was that kind of heterodox economist, I could probably construct a decent argument to the effect that diverting people’s attention away from sensible arguments about institutional constraints toward fanciful ones aout individual psychological kinks was the whole purpose of behavioural economics, as a self-defence mechanism for neoclassical economics.


christian h. 04.01.08 at 11:19 pm

Clearly, all Denton does here is increase the rate of exploitation. It doesn’t need any psychological explanations to see that’s wrong. I guess we’ll have to reserve a lamp-post for him.


Alice 04.02.08 at 12:24 am

The interesting question is whether or not the individual bloggers have any intrinsic value or are only important as an aspect of the Gawker Brand -can they defect and make the same amount (elsewhere or independently?)

Maybe it would be best to publicly announce the editorial budget as a percentage of revenue and then have the bloggers compete for a piece of the total pie


Felix 04.02.08 at 1:49 am

Henry, great piece. But Denton’s not quite the anti-worker monster you’re painting him to be. For one thing, blogs generally (I’m sure CT is no exception) exhibit pageview growth significantly greater than general wage growth. So if Denton wants to keep his wages remotely in line with the prevailing standards, he can’t simply set a PVR and leave it at that: there needs to be some kind of deflator. What’s more, smaller blogs are going to need higher PVRs just in order to be able to attract editorial staff, even if those PVRs are uneconomic to begin with and Denton is essentially investing that money in the hope of future growth.

Denton never really pitched the new compensation system as creating an equitable split of revenue between workers and owner. Rather, he pitched it as creating an equitable split of revenue between the workers. No longer would there be resentments about how much the person next to you was making: they’re now getting exactly the same PVR as you are, it’s a level playing field. The quarterly PVR is then set so that the editorial budget is what the editorial budget is; the payment system acts as a mechanism for fairly distributing that budget.

Now I’m not saying I agree with the system — I don’t, at all — it’s just that I don’t think your criticisms of it are entirely on point. But they’re very interesting, all the same, especially in the light of Marc Andreessen’s blog entry on incentives which is doing the rounds.


sara 04.02.08 at 1:52 am

I’m not a (modern, maths) political economist, but the obvious conclusion is that the site’s quality as news, commentary, or criticism would decline, as paying per the most views inevitably leads to slightly more upscale (hipster?) versions of Fark and Something Awful.

Granted, this may have been Gawker’s business model from the beginning. But it sets a bad precedent for similar enterprises.


Henry 04.02.08 at 2:16 am

Felix, that is fair enough. I did have an extra bit in the original draft which I deleted about how piecework systems are also hard to pull off b/c of internal norms within firms as to who gets what, and perceptions of fairness/unfairness, but decided that my account was getting too complicated (but Miller talks about this too using simple social choice theory – it really is a lovely book). Nor do I think that Denton is a monster, really; just that if he wanted to maintain the kind of system that he wants to maintain, these actions weren’t very wise. I strongly suspect that he will find himself moving willy-nilly to a more conventional system of wages, with some more moderate bonuses on top, precisely because these kinds of political economy of hierarchy questions are so difficult to resolve. Then, I am writing as an academic rather than an entrepreneur, obviously – there may be bits of business practices that are entirely non-transparent to folks like meself.


Henry 04.02.08 at 2:18 am

And for what it’s worth, any increases in pageviews at CT will necessarily be greater than wage growth, given the vagaries of our internal compensation system ;)


dsquared 04.02.08 at 5:16 am

But Denton’s not quite the anti-worker monster you’re painting him to be. For one thing, blogs generally (I’m sure CT is no exception) exhibit pageview growth significantly greater than general wage growth. So if Denton wants to keep his wages remotely in line with the prevailing standards, he can’t simply set a PVR and leave it at that: there needs to be some kind of deflator.

But this is just making the same point that Henry made when he said that “Denton wants to make sure that he captures all the productivity increase for himself”. If he wants to keep wages “in line with prevailing standards”, then he doesn’t really want t pay piecework at all and it’s unsurprising that in the long term his piecework-minus-deflator scheme isn’t going to generate the favourable incentive effects associated with piecework.

(The most famous example of what I’m talking about is the existence of the firm EDS, which exists because IBM tried to cap Ross Perot’s sals commission out of a belief that they could keep his compensation “in line with prevailing standards” but still retain his services.)


Martin Wisse 04.02.08 at 7:26 am

Sow hat’s the problem here? The guy makes nice money, will stand to make less money under the new rules and you need all that explenation to understand why he’s angry?

Do you really need anything more to explain this situation?


Martin Wisse 04.02.08 at 7:28 am

That is, I understand the need for a discipline like economics to be more precise in describing situations like this, but it seems somewhat like arguing about angels dancing on pinheads…


Martin Wisse 04.02.08 at 7:28 am

Also, that gravatar thing sucks donkeys through straws.


Alex 04.02.08 at 10:10 am

But how could Wonkette actually reduce the quality of her blog in order to get her own back on Gawker? It’s not as if there’s much there in the first place…


Great Zamfir 04.02.08 at 10:20 am

Martin, I would say this story is relevant because it is a nice experiment with the kind of performance-based compensations people are trying to implement in many other lines of work too. In most cases, the discussion revolves around the question whether the relevant performance is measurable, but this specific case is interesting because performance measurement is relatively easy, and problems remain.

Sure, the guy is complaining about a ay cut, but there is something interesting about his pay cut: the more succesful he is during three months, the higher his paycut at the end will be, forcing him to be even more succesful in the next three months to receive the same salary.

This shows how ‘performance based’ compensation might result in a very different outcome than you would expect.


John Quiggin 04.02.08 at 11:07 am

Henry, my wages from CT have increased by 87120 per cent since I started blogging here (not counting my contribution to operating costs). Now I find out pageviews have increased even more. Are you tellng me you’ve been ripping me off all along?


Barry 04.02.08 at 12:22 pm

“Sow hat’s the problem here? The guy makes nice money, will stand to make less money under the new rules and you need all that explenation to understand why he’s angry?

Do you really need anything more to explain this situation?”

Posted by Martin Wisse

Actually, yes because most of us are Americans, and most Americans feel that a boss scre*wing over his employees is what Made America Great [1], and that interfering with it is Islamofascocommunosocialfemieconazism.

[1] Sadly, it probably is.


jackd 04.02.08 at 1:36 pm

Great Zamfir (#16) hit the point I was considering. The deflation in PVR payment is exactly counter to the incentive of the PVR itself.

In my own experience, thankfully long, long, ago, the only thing worse than being told the company has decided to lower piecework rates is the insult of a manager saying workers can make up the difference in volume.


Felix 04.02.08 at 2:43 pm

Two points:

1) an increase in pageviews is not necessarily an increase in productivity. Consider, for instance, all the pageviews of stories you wrote months ago. Should they really count towards your productivity this month?

2) The deflation in PVR is not exactly counter to the incentive of the PVR, since the number of bloggers incentivized by the PVR is greater than one. The incentive of the PVR is actually to increase your *proportion* of total pageviews, more than it is to increase the absolute number of total pageviews. In a weird way, it incentivizes bloggers to give their colleagues food poisoning more than it does to increase their own output in line with everybody else.


dsquared 04.02.08 at 5:49 pm

Consider, for instance, all the pageviews of stories you wrote months ago. Should they really count towards your productivity this month?

Obviously yes, unless people don’t serve advertisements to old posts. Otherwise you’re creating an obviously perverse incentive.


Barry 04.02.08 at 6:10 pm

And one obvious reason for people to be viewing old posts is that there is some buzz about the topic and/or the blogger. Example – if I had a blog, and became well-known next year, I’d expect a bunch of visits next year to this year’s posts.

Point two is good – it’s in Nick’s interest (and indirectly in the interests of all bloggers there) for viewers to move around within that site, once they landed there.


Andrew C 04.03.08 at 12:54 pm

But how could Wonkette actually reduce the quality of her blog in order to get her own back on Gawker?

Obviously no more posts about interns having ana! sex with married senators.

Comments on this entry are closed.