cases where structure removed equals process improved
structure v.s. process
Leave a reply
Monthly Archives: November 2005
Niagria
Sun’s new processor offering is kind of interesting. It appears that they decided that the right processor architecture for servers (i.e. machines used by thousands of people at a time) is just different than the right one for more machines used more personally.
If you freeze the machine your using right at this moment and ask it how many threads are standing by waiting to get some processor cycles, i.e. not blocked for any reason other than access to the processor, the answer is almost certainly none, or one. If you designing the chip for that machine it’s worth spending transistors to make one processor run really fast.
If you freeze a reasonably loaded server, say one serving up web pages to a thousand users, with a few middleware applications behind the server, and a database or two behind those. The answer is may well be a few dozen, or possibly even a few hundred. Because this work load is more naturally fragmented your better off spending your transistors on multiple processors.
You might say they asked to question: if we could put a room full of servers on a chip what would that look like? Or maybe it’s the return of Connection Machine. I wonder if Sun’s got a *Lisp implementation?
I spent a lot of my early career working on multiprocess architecture both hardware and software. I became disillusioned. Given Moore’s law they were always only a few years ahead of the curve, which limited your markets and meant once you won customers you lost them a short while later. The action shifted to personal computing and away from servers, where the problems were easier to map onto the hardware. Mapping customer problems onto parallel architectures was sometimes easy; but more typically it tended to require too many clever engineering hours to be worth it. The programming languages, compilers, and operating system architectures that would reduce the how clever your engineers needed to be tended to be immature and the market wasn’t deep enough to spin up the network effects to fix that. That was the first time I began to appreciate the value of moving where the network effect current is fast.
Two things really triggered my exit from the multiprocessor industry. One was the Macintosh, it was just so cool and it had so many obvious opportunities to do neat things. The other was the death of the line printer. We used to have these line printers, and they would have one print head for every column on the page. And then we bought a new printer and it only had one print head that went back and forth. It was much faster. I thought – Yeah, if you can’t even make the case for multiprocessing in a task so transparently parallel in the physical world how likely is it that this whole multiprocessor thing is going to work out?
I wonder if it will be different this time? Three things have changed. Much of the action has shifted back to the server side. The total size of the market is just amazingly larger – which can feed the network effects needed to build the tools. Our talent at this kind of stuff is much deeper and broader. So it seems like it might.
But it’s daring to decide that the processor industry has a viable niche where a new species can thrive.
Developer Network Pricing: Now Free!
eBay has decided to give access to their developer network and traffic thru their web APIs away for free. That’s almost two distinct decisions.
According to News.com, up until today developers paid eBay between $1.25 to $2.90 per 1,000 items listed and an annual fee of $500. (*)
They are leaving a lot of money on the table, an that doesn’t happen easily, so I’m always interested by firms that decide to give away their developer networks and or their APIs because internally it makes the developer network group dependent on internal politics for their funding. Which means that the firm must have a strong model for what value they get back from these activities. It doesn’t matter what you call them: open APIs, developer networks, etc. etc. the problem remains. Most firms won’t build these if they can’t clearly comprehend why.
Here’s a start at a list of why firms give this stuff away.
The simplest case to make arises you can draw a clear line from the usage of the open API to revenue in the heart of the firm’s business.
- for example eBay or Amazon can give away open API access if drives more transactions thru their marketplace,
- for example a programable logic array vendor can give away designs tools if that leads to increased component sales,
- for example Google can give away a mapping widget if they are confident of their ablity to float ads or capture traffic from it, or
- for example they can give away website analytics if it tends leads to increased add revenue.
You might do it to solve a search problem. Firms typically have lots of options for what to do with the stuff in their portfolio. Searching for
the best things to do is hard and risk.
- Part of what makes it hard is that firms are don’t know what problems customers have, only customers knowthat. So a toolkit given to customers who are close to real problems can overcome that problem. How was google to know that people needed a mapping widget to display cheap airline fares?
- Developers will take risks that firms won’t. For example a large firm will insist on creating a web site that is scalable, internationalized, etc. A small firm will defer that; and thus learn faster. Small firms are less regulated than large ones.
- All firms have cultures the define what they consider to be important. One firm may care about safety. Another about ad dollars. Another about cool user interface. These core values make it impossible for the firm to discover innovations that involve a different mix of core values.
You might do it to create a network of complements around your offering.
- Complements raise the percieved value of your offering. This is why people sometime switch from the more robust FreeBSD to the more richly complemented Linux.
- Complments can drive upgrade from version to version (if that’s your revenue model). For example there comes a time when I must upgrade my Mac’s OS to gain access to current versions of other software I use.
- Complements make your offering much more sticky because when the customer wants to switch you have to coordinate switching all the complementary products you use.
It can be a useful part of your pricing strategy:
- As a kind of free sample.
- As a kind of lead generator. Developers who adopt your open APIs are likely cantidates for adopting more pricy offerings.
- As a way to get the long tail of users unwilling to pay drawn into your network. This is particularly key if gain some positive value for your network from every participant.
Then there are the reasons that arise from a firms desire to shape the market dynamics around it.
- An open API can be used to set a standard, and having the industry standard flow from your firm maybe in your best interest.
- An open API can draw in a large number of smaller players, who’s cost of sale you couldn’t reasonably bear, and temper the power of the large players who’s cost of sale you can afford.
It maybe in your interest to commoditize an adjacent market so your core offering remains were the money is made in the value chain. For example one model of open source is that software developers would prefer to be paid for their work and eroding the power of IP rights holders to charge pushes the money away from them and toward the coders.
The motivation for any particular developer network is probably an 80/20 mix. With 80% of the motivation coming from just a few of these, and the remaining 20% comming from the rest. I suspect that over time the motivations shift. I’ll leave it to the reader to puzzle about what shifting motivations drove the eBay change in pricing.
Vendor Loyality and the Genetic Imperative
I am a sinner! I am guilty. Yes, I too have pulled this stunt: baby pictures in your messaging. But, but, I only use the happy baby pictures to extract from my audience subliminal love for my message.
This vendor is subliminally attempting to trigger my love, my loyality. He want’s a hug. The email this image appeared in reminds me that I haven’t dropped by and used his service for over a year. This is much more evil than my happy baby pictures.
That was quick
Ian Holsman writes:
Google has just released a free service to do web site analytics which is/was their ‘urchin’ product. sorry all you startup’s who were in that space?.. maybe next time ;(
It wasn’t too long ago that developers and VC rushed into the Internet to escape this pattern on the desktop, where Microsoft methodically vacuums up any idea likely to have wide applicability – turning them into features of one of their hubs.
It’s very common for hubs to absorb adjacent universal functions. Three forces, at least, drive that: appetite for features, keeping your complements commoditized, and scale efficiencies. In the Microsoft story one of them is/was the need to find new features that can drive upgrades. Another, also seen in the Microsoft story, is the strategic necessity of keeping competing hubs from emerging in their market – for example if they hadn’t taken the word processing network away from competitors then they would have been forced to negotiate with a powerful partner rather than lead of weak developers (“chasing tail lights”). Finally there are pure production and demand side scale advantages.
Of course in some cases the story doesn’t demand a complex model like that. Adding analytics to your ad network is just a feature, like adding footnoting to your word processor.
I liked a line I read recently about how so many ventures in there days are “built to flip,” but that the speaker’s firm was being “built to last.” When strong hubs emerge in a market, like Google, Yahoo, Amazon, eBay, etc. flipping becomes a valid move in the game. There was a time when one of the more dominate models for a venture was to flip it into Microsoft, and while that’s still a common model it’s not as rampant on the land as it used to be. In those days the Microsoft developer network could be productively viewed as an extension of their acquisitions or R&D strategy.
When this happens in a market it has two consequences. The ventures become much more winner take all games; since there is rarely more than one place to flip a particular effort into. That kind of risk drives developers to look for other markets to play in. The customer also changes; if you know that your firm’s future is very likely to lead to flipping into Yahoo, or Microsoft, then they become the customer and the day to day customers are no longer job 1. That creates lousy dynamics for seeking really useful innovations – you’re now innovating for the hub, not for the user.
Club Pricing
Paul English is peeved about what a pain in the neck it is to buy a membership in a health club. He asks “Why the sleaze?”
Here are some theories:
- Low barrier to entry means too many health clubs; i.e. excess supply, leading to sleazy marketing practices as they desperately attempt to survive.
- Lack of consumer protection laws or enforcement removes the only effective negative feedback on the sleaze.
- Testosterone and steroids.
- The subscription pricing models (i.e. lock-in and upfront discounting) is sleazy out of the gate, it’s all down hill from there.
- Clubs that adopt aggressive value pricing are more sustainable than those that don’t. The more aggressive the more sleazy.
- Matching price/user to cost/user is impossible, all attempts to do so look and often are sleazy.
Clubs are the text book response to the standard list of problems with of public goods, e.g. overcrowding, under provisioning, free-riding, etc. It’s fascinating that in this example the club doesn’t resolve those problems; it just reframes them. The presumption is that once you are allowed into the club you will find a world where facilities and services are abundant and collegial; i.e. you will enter a world where the facilities are now a public good. You won’t be excluded and there won’t be rivalry. The water in the pool will be warm and you won’t have to double up in the lanes when you swim. Piles of towels will be close at hand. The exercise equipment will be standing by and dependible when you need it.
Most of the health clubs I’ve experienced fail to achieve the eden like fantasy. They are under provisioned and overcrowded at the hours when I would show up. Interestingly the lanes in the pool would be occupied by people who used them for hours every day, in effect free-riding on the membership contributions of people like me.
These clubs have terrible pricing problems. They would like to charge people for the value they extract from their facilities, including the value of reducing their guilt by having a membership even if they don’t use it. Of course they don’t mind if they over charge. The more they can push their prices up toward that goal the better the facilities they can provide will be. That’s all reasonably virtuous.
I have a bad feeling that the distribution of load imposed by the customers is power-law in shape; but the willingness to pay isn’t. Which is why on the high end you get the heavy user free-loading illustrated above. On the low end you get a long tail of users who are charged far more than they will ever consume in services. The entire thing’s a mess.
I have noticed that some very high end hotels have health clubs that don’t suffer from these problems. They, presumably, are subsidized by the hotel’s guests. In a few cases they even sell a reasonably priced membership in a transparent fashion.
The YMCA is a good special case. As a not for profit they can do their value pricing by using needs based analysis. Their prices are clearly stated on a sheet of paper at the front counter. At the bottom of the sheet it invites you to speak with them if you can’t afford the price they are offering. At the same time solicit donations from the better off members of their community. It’s interesting how this doesn’t suffer from the sleaze problem. It also means that when you gain entrance to the club and the public good isn’t as wonderful as you might hope your reaction isn’t that you have been mislead, but rather that a group of good people are doing the best they can with limited resources. Which, if your well off, might lead you to donate.
At the same time the Y’s model for how to fund the club doesn’t resolve some of the problems outlined above. They still have some members who draw off 100 times more services than others – i.e. free loaders. And the club are always a bit under provisioned, the water a bit chilly, the equipment a a bit run down. The community is generally very convivial; particularly if it makes you happy to see hordes of elderly and children using the club around you.
The private health clubs have eroded some of the communities common cause around the Y. That has weakened the Y as an institution. For example the Y wasn’t able to raise money to build a new branch in downtown Boston recently.
Universal Demand for Quality
I’ve been reading some books about professionalism. The one I was reading yesterday touched on an interesting model that I’m not quite sure what to make of, but it certainly caught my fancy.
Some commodities have universal demand; i.e. everybody wants some. Some examples: food, education, knowledge, safety, health, mobility, conflict resolution. States naturally are drawn into providing a regulatory function for these industries. A state that fails in these areas finds its legitimacy at risk. Their universal demand assures a strong signal from the citizens to the state, particularly in any functional democratic state.
Meanwhile, universal demand tends to attract numerous suppliers; and in the absence of barrier to entry too many suppliers. Which will lead in short order to market failure if the quality of the goods supplied is hard to measure dependably. The market fails because the horde of suppliers furiously underbid each other until they can’t make a reasonable living, which drives all the competent suppliers to seek other work.
The lack of clear quality measures leads the substitution of alternate sources of legitmacy: pomp, pompus attitude, parasiting on other sources of authority, advertising, character defamation. (A point which deserves a blog posting of it’s own, but since that’s unlikely I’ll toss in this marvelous line. When this happens you see a pattern: consumers hold the trade in very low esteem but hold their personal practitioner in the highest regard. Where have I heard that before?)
It is practically impossible for most buyers to evaluate the quality of what they are buying. If you can’t tell from the plate on the table in the resturant if the kitchen is or isn’t a public health nightmare there is no chance you can evaluate the quality of your teachers, lawyers, groceries, or the city’s levee.
(Oh no, another aside: The library I was reading this book in has taken to using the fire alarm to annouce that the library is closing. It’s a nightmare waiting to happen – the fire that breaks out at 20 minutes before closing time.)
So that’s the story. A commodity with strong demand whose quality isn’t transparently obvious can easily engage in a rush for the bottom, a market failure. If the demand is universal the state will find the pressure to respond irresistible. And so the state will step in to regulate.
As early as the 16th century some European states established regulatory mechanisms for medical providers. Now that’s a great example because it looks to me like those states picked, more or less at random, one class of medical hucksters declared them legit, and declared the others ill-legit. They had to do something.
What I find thought provoking is how granting the state license (the franchise, the monopoly) to one group is a new kind of standards making I’d not recognized before. In the presence of 16th century medical science (i.e. a something totally bogus) and complete market failure (i.e. doctors and barbers sharing the same wages) the state has a chicken and egg problem. No quality, and no market. By tagging one group as responsible it solves the market failure. Wait a few centuries with luck the might science emerges.
If the practitioners don’t capturing the regulator and the regulator keeps demanding that the practitioners address the quality problem this can work. There is some hint that is exactly what happened with the medical profession. That time and time again the profession failed to provide reasonably quality; for example Doctors were very slow to adopt ideas about public health, hygiene, etc. If you want to be nice you could say they were very loyal to their professional practices. Forces would come to bear that would force them to change; in the absence of the professional monopoly there wouldn’t have been anything for those forces to bear down upon.
Identity Blackmail
Credit rating firms are the premiere example of an identity business in my gossip model of identity. They aggregate gossip about citizens from institutions that have active relationships with them know them. They then sell models of those citizens to institutions that would lack a model but need one to reduce risk. You and me, those who are being modeled are typically not their customers, we lack any relationship with these gossip firms. They have relationships with their suppliers, firms with models, and their demand comes from firms without models.
That’s been changing as consumer protection laws have begun to force the credit rating firms to develop a relationship with the consumer. That’s turned out to be profitable.
These gossip firms aren’t limited to just credit rating. Some of them will talk about credentials – criminal, academic, licensing. Some of them do medical records. I assume there are ones for insurance and physical location, etc. etc.
The Internet is beginning to provide an interesting new source of supply for the gossip companies and new business models for building them. You can aggregate a lot of information about some people using just a search engine; and who knows maybe it’s higher or lower quality than the information a more classic background checking firm could get you. The social networking sites are kind of gossip firm – with much smaller suppliers and customers than the traditional credit checking firm.
A friend asked recently how he could fix the bogus links that come up first on Google when you enter his name. These were articles from a local newspaper full of inaccuracies. Since there is no consumer protection laws around google’s role as a gossip intermediary my answer – a somewhat more nuanced version of ‘get better fresher gossip’ – wasn’t particularly helpful.
Today another friend sent me a link to a web site with a model of everybody. They have scraped the web trying to find each and every one of us, and then populated their model with what they found. Here’s what you get if you look up my name.
Recalling that the credit check firms where forced into the discovered that it can be profitable to create a relationship with the people they are modeling. Recall that the social networking firms (orkut, linkedin, friendster, etc. etc.) start right out the gate by creating a relationship with the people they are modeling.
I’m amused to notice that the folks at this place have a button that allows me to claim my page. Since much of what they collect is woefully incomplete and full of errors this button looks a lot to me like blackmail. The first time I saw that “claim your page” technique was at blogshares – a delightfully silly game built around how many links your blog has. Technorati has a similar device. So it’s not always blackmail; though it does always have just a hint of something odd. That’s the nature of gossip.
Thanks for two typos so far.
Building Houses on Sand
Philip Jacob worries, and rightly so, if this current enthusiasm for building websites that rollup bits and pieces isn’t, well, insanely risky.
“…would you incorporate a library in your application that was licensed under terms like this?
- Required you to read and understand several pages of legalese
- Is free right now although you might be charged for it at some point in the future at a price that you cannot negotiate in advance
- Can stop working at any time for any reason without notice
- Can undergo functional changes at any time without notice
- Can be rate-limited at any time without notice
- Does not have any service level agreement
- Places restrictions on the data you generate using the library, some of which are bizarrely techno-legalistic and open to interpretation”
The licenses of all these fun web APIs are mavalously diverse. It is clear that each firm’s lawyer plays the role of genetic engineer gene splicing together a handful of licenses to create yet another mutation. While in long run (after you’re dead) the ecology will put most of these out of their misery. License interop is hard. This is another example of “standards as a substitute for lawyers.” Phil would like to have some standard licenses; but
We know how this plays out. First we get a bloom of diversity. Then we get consolidation. In the end we get a power-law distribution across some number of licenses. In the final end game the elite members of that distribution engage in a long standards battle who’s outcome can not be predicted – they might make peace, they might split into seperate markets, one might win, etc.
In early phases the smart players move quickly to capture market share that will provide key negotiating power in the later stages. Some of the early movers have to be profitable, others don’t. Some players value a strong position in the future more highly than others. Some players discount risk of having to switching later.
There are adjacent disputed territories to consider. Consider four of the larger players. The communication companies (though they haven’t fully accepted their fate yet) lock in users with physical networks and charge for bandwidth by the month. Companies like AOL, Yahoo, Google charge via advertising and lock-in by capturing a large distribution bottleneck. Platform vendors charge for a stream of updates and lock in via APIs and licenses. Players like the FSF or the ASF are sort of inside out with their focus on freeing developers (their principle consituency) from the licensing/API lock-in points.
All this is just part of the passion play around the emerging Internet OS, and to think I used to believe this was all kind of simple.
What do I need line for, when I have color?
I am not worthy! Peter Davis attempts the impossible; to reduce thousands of man years of work on PR agendas, industry politics, entrepreneurial hope, market share machinations, expert puzzle solving, and late night enthusiasms into a simple block diagram! Maybe version 1.1 could show the number of email messages, frequent flyer points, and billable legal hours. Oh, possible he could hint a the IPR dispute density with drop shadows?