But that’s not exceptional.   Why this happens is full of fascination.  Metrics management encourages unethical behavior.  In some cases this is intentional but I think that’s rare.

What happens is easy to explain if we set aside the question of ethical behavior.

The manager’s problem is always how to coordinate behavior toward his goals.  If the existing behaviors are fine, and often they are, he want’s to add additional behaviors.  Metrics management is tool for doing just that.  He creates metrics that highlight the desired behaviors and then works to bend the curve.  Just making the metrics visible is usually sufficient; but people tend to escalate so it’s common to pile on some carrots and sticks.

Lots of us who have used these methods have found them to be marvelously effective.  Early in my work life, for example, I put a chart on my door showing the number of open bugs in the large software system we were building.  The line wobbled along, moving up and down, over a period of maybe a year.  That was, pretty much all I had to do.  In the following six months the curve steady declined until it finally crossed zero.   Managers, engineers, customers suddenly became focused on that line.  They starting asking about stupid little meaningly wiggles in the trend.  It was magical.  When we finally crossed zero parties and tee shirts emerged.

Other things suffered.  And that’s the thing.

Or no, it’s not.  Presumably as things suffer the natural response would be to adjust.  And yet people don’t.

There is a weird pattern here.  Any organization is skilled at some things, call those A.  If we had metrics for those it’s scores would be pretty good.  But we typically do not have metrics for these.  Not because we are incompetent managers, but because the behaviors that achieve A are so deeply embedded into the organization that they do not require close supervision.  These behaviors are sustained as norms, manners, etc.  These behaviors are so fundamental that they are part of the organizations social fabric.  The social contract it has made with it’s self.  This fabric is slow to change.

So when the metrics management card is played into the game, in service of adding or improving a set of behaviors, cal them B, a very odd thing happens.  Of course on thing that happens is that B behaviors increase.  If they didn’t managers wouldn’t use this trick.   For a period of time the overall score A+B rises.  The A behavior score does not decline (at least not much) and the B behavior score rises.

But in time the A behaviors start to decline.  This alone is an interesting pattern and I don’t think I can recall any instance of metric management that didn’t manifest this pattern.  At first, and surprisingly rapidly, everything gets better.  This a quite delightful experience.   It’s positively addictive.   Later A scores decline, but since your not measuring them this tends to go unnoticed.

Except it doesn’t go unnoticed.  Individuals in a social fabric are extremely aware of the social contract.  If only because it’s enforcing feedback look is implemented thru millions of tiny social cues.

What appears to happen as metric management moves into a organization is that it displaces the existing social contract.   This triggers a lot of cynicism.  Which in turn undermines ethical behavior.

Businesses adapt the ratio between guard labor v.s. productive labor. That ratio varies across firms within industries, from one industry to another, and inside of firms from on department to another. Presumably there is a great deal of peering over the walls. A lot of monkey see monkey do as people search for a different balance. For example if the neighbor organization appears more successful and they don’t spend time on daily progress reporting and your organization does then you might be tempted to jigger things; reducing the ratio of guard to productive labor.

The work done by labor shifts. It can shift into automation or codified processes and procedures. If the neighbor is using an elegant status board your organization might be tempted try that out. In effect shifting guard labor into automation. Processes and procedures are a capital investment. How you depreciate and retire them feeds directly into organizational agility. If a neighboring organization appears to be successful by setting aside their rule book and replacing it with a charismatic manager then you might try that. In effect shifting automation into guard labor.

If you cast this in an exaggerated light you’ll notice plenty of opportunity for polarization. For example guard labor may fall into the bad habit of talking up how lazy the productive labor is; not because they dislike the productive labor but because it helps to justify their role (talking their book). Same thing happens for al the pairs; e.g. X v.s. automation and other roles like customers and suppliers. Most organization players flirt with the guilty pleasure kind of such exaggerated behaviors. It’s worth noting that there is plenty of opportunity for similar polarizing behaviors to be found the variability across departments, between firms, etc.

Automation (or processed and procedures) means that productive labor can be highly guarded even in the absence of guard labor. The assembly line worker’s role is extremely guarded, by the machines that surround him. In that scenario the guard labor ratio is low, with say only one supervisor for a a dozen production workers. So how guarded work is somewhat independent of how much guard labor is involved. Mechanical turk workers might be a example of an outlier in that space – highly guarded work with very minimal guard labor.

It is a challenge to label the axis for how guarded labor is. One extreme might be labeled slavery, bound labor. At the other end we might expect to find labor that is enjoying very high values of achieving the people’s core concerns (appreciation, autonomy, affiliation, …).  There are plenty of things that can bind labor: salary; stock options; roots in the firm or locality; specialized skills; lack of other options. In the Americas importing slaves from Africa was more effective v.s. enslaving the natives, I presume because natives had the option of running away.

Like all costs the cost of guarding can internal or externally born. One point the paper makes is that unemployment acts as a kind of systemic guard. Obviously unemployment insurance acts to temper that effect. If incidences of unemployment tend to be short in duration, that also tempers the effect. The paper suggests that Italy, with high unemployment and low guard labor ratio, might be an example of an economy where the cost of guarding has been externalized. In this mindset taxing firms that tend to have high labor turn over is a Pigovian or pollution tax and helps to encourage the guarding to move in house where it will be on the books.

This posting about slavery is interesting to mix up with the guard labor meme. It’s all a bit rough but there are two rolls of the economic dice mentioned there that encourage the rise of slavery. Let’s decide that what we mean by slavery is state sanctioned and enforced slavery. Not that the other cases aren’t interesting.

The first scenario is the case where owners have land, but labor is scarce. The scarcity of labor means that the cost of owning slaves is lower than the cost of paying them a salary on the open labor market.

The second scenario is high demand for some good. I don’t entirely understand this suggestion; but i find it plausible. If I replace high demand with high volume it begins to make sense. If some labor intensive good is produced at very high volume an large industry will emerge and scaling effects will kick in. That will trigger the search for both the balance of guard to productive labor, the codification of process. In time the nature of the productive labor will become increasingly guarded. The guard labor ratio will rise – the absence of either automation, or technological schemes for assuring the productive labor adheres to the processes. In the extreme case that high ratio will lead to introducing slavery.

Of course you’d also need to be able to implement the bondage. So if the labor can run away, as in the early American colonies, it maybe a lot harder – even with the help of the state and cultural sanction and enforcement. Labor can also run away if you have a diverse economy; so it would seem to me that both these scenarios are much more likely and possibly even require homogeneity – a company town.

The first scenario suggest that in any case where labor becomes scarce the incidence of binding devices will rise. So when labor is scarce during the second world war and wages are capped employers introduce health insurance, which has turned out to be a very effective binding device. It is exaggerated to suggest most of these are equivalent to slavery, but it’s interesting to float the analogy.

Some commodities suffer from an imbalance, high demand and low quality of supply.  The skill of how to get the girl is one, as are cures for cancer, weight loss programs, or how to close a sale.  There is a kind of evolutionary arguement to be made that in all these cases if a high quality solution were to emerge the other side would come under powerful evolutionary pressure to discover a counter measure.   One reason the hucksters thrive in these markets by virtue of the plausable premise that it just might be some secret high quality trick to it.   Cancer?  Positive attitutde.  Weight loss?  Bacon!  Close a sale?  ”Would you like it in blue or grey?”  Get the girl? Demonstrate value and  play hard to get.  Humans are a mess, the ultimate rube goldberg device, so these all work.  Sometimes.

You could write a book like Neil Strauss’s The Game about any one of these markets the exhibit high demand and an unlimited supply of low quality goods.  And in each case you’d get the same assortment of characters; the desperate, the needy, the clueless, the hucksters, and the occational guys with talent.   You’d also get that delightful pattern, common on the internet, of groups of common cause forming. Random samples of people who share the problem at hand who gather and toss about ideas about what works and what doesn’t work.

The nature of such groups can cut across a wide spectrum from cheerful good fun, thru wholesome, into vile, and unto distructive cultism.  In the venn diagram of what kind of book The Game is one bubble should be about the transition of one such community thru all those stages.  At the beginning we have a bunch of dweebish shy disfunctional guys who are teaching each other to take a bath, wear snappier cloths, how to approach a stranger, how to make small talk, how to avoid wearing out your welcome.  At the end we have power hungry entrepeurs pulling down vast sums of money to teach this demographic the skill of being assholes (see photo of author and his teacher) and how best to apply their new found skills – approach women in quantity.

The venn diagram of what kind of book this is would include quite a few more bubbles.

This is certainly a book about cults.  And I might add it to the small pile of my favorites.  It’s a rare example of a “I was a cult victum” narrative where the author is not entirely angry, alienated, and damaged at the end.  That said I suspect there is more of that then he is letting on.

This is certainly travel narrative of that fun kind: fool goes to strange and exotic foreign land where he behaves like an idiot and makes a long series of very bad choices.  As readers we get a continual perverse frisson from that.  We regularly roll our eyes, gasp in disbelief, and take comfort in the fact we wouldn’t be such a bozo.  By way of example at one point he, as instructed, picks up a set of thick acupuncture needles and shows up at he dissheveled home of an amazingly dysfunctional celebraty where she alternately sticks him and runs out for junk food.  And that’s only an example!

It is also a fine example of the classic story of hero leaves home, has adventures, return home wiser.  But oh our hero is flawed, which makes us sad.

It is also a comedy, we know because it ends romantically.  But then is is also a tragedy, since many people die – well they don’t necessarilly die but there is a souless cult leader with his nest of scary of zombies left unresolved at the end.

It has that nerd, fantasy fiction, geeky element where in you learn a secret language.  Not Kilingon.  I was reminded of that fun book Edge City where you can learn bits of the secret language of Real Estate developers.  For example here we learn the term “Chick Crack,” i.e. those little personality surveys found at the back of women’s magazines.  There are plenty more.

I recomend this book for all that.  Who doesn’t like a book about men behaving badly.  It’s expensive, but if you get it from your local library you get a kind of director’s edition.  Since at least one sad sweet shy dweeb will have selectively underlined portions in the hope of treating his problem.

(I have a bad feeling this post is going to attract a lot of spam.)

UPDATE (December 19, 2009): Peter Staats, in the comments, suggested that belief in anthropogenic global warming is entrenched among scientists and will disappear as the older generation dies (citing Planck, whose point is also made in Thomas Kuhn’s Structure of Scientific Revolutions). I responded that I thought he has it backwards–that AGW has become more and more supported, and the holdouts tend to be older, as some of the data about the anti-AGW organizations above already suggested. So I tested our respective hypotheses against Jim Prall’s data, for IPCC WG1 scientists vs. the signatories of the AGW-skeptical documents. I looked at the average year of the last academic degree awarded, first for those with citation counts for their fourth-most-cited paper >= 200, then, since that was such a small sample for the climate skeptics, for citation counts >= 100, and then for all the 623 IPCC WG1 scientists vs. the 469 signatories of AGW-skeptical documents. Here are the results:

Citation counts of 4th-most-cited >= 200:
IPCC WG1: N=83, 12 w/o year, N=71, average year of last degree = 1981
Skeptics: N=13, 4 w/o year, N=9, average year of last degree = 1965

Citations counts of 4th-most-cited >=100:
IPCC WG1: N=201, 51 w/o year, N=150, average year of last degree = 1983
Skeptics: N=38, 15 w/o year, N=23, average year of last degree = 1968

All IPCC WG1 vs. AGW-skeptical document signers:
IPCC WG1: N=623, 208 w/o year, N=415, average year of last degree = 1989
Skeptics: N=469, 346 w/o year, N=123, average year of last degree = 1973

In short, the skeptics are 15 years older than their opponents.

My mind is settled on this whole climate issue.  But, it would be fun to have a list of other conflicts with a strong generational component.  For example, I think I’ve written about how I think Microsoft is caught on the old side of the swing back to data center computing and it’s associate control of the customer and distribution channels.  I suspect there is one in the Business schools around platform based business architectures.

In a somewhat related bit, I’m subscribed to various pages at wikipedia.  Apparently there is some kind of lower life form that draws satisfaction from swooping into pages and adding “citation needed” after the first period, and so naturally there are mechanisms that chase after this to remove them.   This silly activity happened on one page I was watching the first sentence read “Given two similar rewards humans show a preference for one that arrives sooner rather than later.”

And so, I was delighted to stumble upon a possible citation for that.  Via Crooked Timber, quoting Richard Tol and his co-authors:

Estimates of utility discount rates for individuals are almost always positive – an estimate of 1.5% is considered plausible for the UK for instance (HMTreasury, 2003) – for the simple reason that humans prefer good things to come earlier rather than later. Given the inevitability of death for individuals, a preference for benefits to accrue earlier rather than later is entirely sensible.

John Quiggin adds:

We can sharpen this up a bit by observing that the average annual mortality probability for adults is around 1.5 per cent, suggesting that this factor alone is sufficient to explain positive time preference.

That whole post is thought provoking but what has been stewing in my head since reading it is how immortals might manifest a very different discount rate than we mortals do.  The planet, society, nations, and many corporations act as if they are immortal.    This leads to a pervasive mismatch in the discount rate between an group and it’s members.  That’s the tension of the prisoner’s dilemma, the gang would prefer that the game continue.

He labels the first of these “perversity.” Here in while reactionary pretends supports the goal he then goes on to explain that efforts toward that end are certain to backfire. Efforts to improve health care? Such efforts will decrease health care! Universal schooling? Such efforts will lead to wide spread idiocy. Do-gooders make things worse. The audacity of this argument is breath taking. But look at the record! How that French Revolution turn out?

Hirschman points out that observers of the French Revolution quickly deployed this argument. Even before the it all went to hell in a hand basket. Edmund Burke in particular used this perverse argument, and later when it things got ugly he got a lot of credit for being so insightful. So did Burke invent this technique? Hirschman argues that no, Burke was mimicking newly popular argument with a similar structure that had recently arisen in the circles he ran in. I.e. the hypothesis of Adam Smith. Aka, the Invisible Hand. This takes my breath away!

The invisible hand is a perverse argument. But in this case bad actions (individual greed, personal vices, and self interest) have the unintended consequence of creating a vibrant national economy. It’s as if God in his infinite wisdom had sus’d out how to turn his flock of sinners into something constructive. Smith might have given credit to divine providence but choose to give the credit to more amorphous but still spirtual invisible hand. Many of Smith’s readers saw right thru that. Particularly all those commercial actors looking to get the church off their case.

I can’t seem to stop chewing on this. It goes in all kinds of directions.

There are numerous systems where actions sum up to something surprising. I can’t believe that I hadn’t noticed how Evolution and the Invisible Hand are both theories of a kind. In evolution the bad (dyslexic genes, mutations, death, mindless long time) that shapes inconceivably marvelous species. There is, it appears an entire class of theories where acts of ethical kind sum to results of the opposite kind. God works in mysterious ways.

I am enjoying this book.

Social networks are a particularly interesting test case for looking at issues of the multihoming since of course they are where you make you home.  I have account memberships in about a half dozen different social networking sites; but I don’t actually participate in any of them.  Though that all depends on your definition of social networking site.  If your more generous in your definition, including say all the on-line forums and mailing lists that include a social (v.s. purely on-topic) component then the number of sites I have accounts at explodes.  A quick review of my password wallet suggests the number gets up toward a hundred; the phrase ‘a gross’ seems useful at this point.  An then some percentage of the blogs I read have a social (or community) subtext.

Some of these places are quite social.  The shaving and diet forums for example. The Oil Drum and Crooked Timber are two nice examples of blogs that sustain a community around them.  Others are semi-social; the one for my PDA for example.  It’s worth noting in passing that the social can make it a bit tough to keep the sites useful for their nominal on-topic purpose.

Where you sit changes how you look at the question of multihoming and social networking sites.  If you have a large stake in one; owning LinkedIn for example but even if you have invested a lot of your social energy into a particular one your profession for example, then you are likely to be interested in ways of reducing the degree of multihoming.  There is certainly plenty of literature on how to execute on that.

Lots of people interested in knowledge flows have noticed that that individuals that cross between two social/professional networks often account for critical knowledge transfers.  So if your interested in encouraging that kind of thing then you might be interested in how to manage and enable increased multihoming.  I don’t think I’ve ever worked for employer who failed to consciously, though rarely conscientiously, encourage a modicum of that kind of thing.

Multiple social networks create some diversification, which in turn can be a buffer against various risks.  Two risks bear mentioning.  If a social network goes bad having other networks enables members to exit, but also it enable them to be critical and that critique can be key to fixing what going wrong.  Having multiple social networks also allows members to take risks, not just of criticism, but also to take risks that may do irredeemable damage to their reputation; such risks are much harder to take if there no other network to retreat into.

None of this helps to puzzle out the question of exactly how many of my ‘professional’ associations I should enumerate on this form.  I’m sorely tempted to enumerate the complete list of all the on-line forums I’m a member of to which I both feel some loyalty and have any overlap with my employer’s vast range of activities.  Just for fun.  Oh, but curiously I appear to be in a job category where they decided to wave the requirement.  Well golly, now my feelings are hurt – they don’t seem to care if I’m a two timing disloyal abuser of the brand!

Back in the 1970s and 80s I spent some time associated with the knowledge engineering crowd. That community labored to build a class of software systems, expert systems, that could perform as well as an expert in this or that narrow technical domain. For example let’s say you had a huge expensive chemical plant. In that plant you’d find a guy, call him Joe, and Joe knew how to start it up. It would take a few days to get the thing running and Joe was the guy who knew how to do it. Did I mention? Joe is retiring next year. In the knowledge engineering crowd approach to this problem was to see if you could extract from Joe, via observation, interviews, what ever, a codification of his knowledge. In the AI/expert systems branch of knowledge systems the idea was to code it up software. The unit for such encoding was rule; i.e. Given that the pump in the basement of building 7 is making that funny noise it sometimes makes delay starting the boiler in until the noise stops.

This turns out to be much harder than it looks. Which we probably knew going into it, since back in the 70s it became common to observe that it took experts about a decade to become competent. That if you modeled the scale of their rule set you can then say that they learned a new rule at the rate of about one or two a day over that decade. It is unlikely you can pull the rules back out of Joe’s head much faster than that. Joe can’t just rattle off his rule set as if you were down loading some file. For Joe these rules are intuitive. I like to say they are compiled in. He doesn’t think thru why the boiler’s start up should be delayed, and in fact he may not even be able to tell you that he’s waiting or the pump to stop making the funny noise. At least he can’t tell you without a effortless moment of introspection.

I was reminded of all that as I read the delightful chapter on the “culture school” of strategic management. The culture school had a few years of popularity when the Japanese cars caught the attention of the B-school crowd. Any number of them up and ran off to Japan and for many of them it was the first time they had seen a radically different culture, i.e. Japanese culture. So a favorite theory what the strategic magic Toyota had that GM didn’t was culture.

But what is culture? It’s unlikely that GM could have been saved by introducing underwear vending machines; but would moving all their suppliers into a dense single city have helped? Culture is like expert knowledge, decompiling it is very hard. If you stop a member of a culture in the midst of some activity and demand “so, why do you do that?” What is the functional value of sleeping on the train? Standing up?

I was delighted by the answer they suggest in Strategy Safari, i.e. that Culture is exactly that which you can’t explain; i.e. it is the expert knowledge which you haven’t codified and made rational. Which of course makes it a bit difficult to manage. If your faced with a competitor who’s advantage over you is cultural the challenge is convert culture into codified knowledge. That’s hard, like getting Joe to mention that thing about the funny noise. Machine that changed the world is a good attempt at that for the Toyota example.
Strategic Safari has a nice framing of why culture is valuable. Firms have unique resources; e.g. capital, location, skills, property, talent. Some of these unique resources are particularly unique because their competitors can’t imitate them; e.g. these resources are valuable, hard to substitute for, and rare. When Steve Job’s brings a few other CEOs on stage during his MacWorld keynotes he is signaling just that: Apple has Steve and as you can see these other guys ain’t Steve. Microsoft has HotMail, but competitors like Google and Yahoo are able to, over time, imitate it and build a substitue.

Knowledge resources are particularly easy to imitate, but only if you can codify it. The harder a company’s knowledge assets are to codify, aka cultural, the more likely they can actually provide a strategic advantage. A valuable culture will be hard to codify. They also have nice five step recipe for how to kill a culture (useful when faced with a dysfunctional culture):

1. Manage the Bottom Line, no actions that can’t rationally explain their benefits.
2. Plan every action, avoid spontaneity and thus learning.
3. Move managers around preventing domain expertise from displacing their managerial skills.
4. Always objective, aka portfolio management.
5. Always use recipes with five steps.

Discussing with a friend how culture is the asset you can’t rationalized he mused that it sounds like the problem many super heroes have e.g. that their super power is accessible only via some irrational pathway. We chortled at the idea of a fantastic four of strategic marketing. The Johnny Storm of PR; the Incredible Hulk of closing; Mr. Fantastic of discriminatory pricing; and the Invisible Woman of customer support.

“Tyranny consists of the desire of universal power beyond its scope.”

One of the nice things about having a blog is that you can spit out those damn brainstorms before they do too much damage to your equilibrium or worse or are extinguished by your daily life.

I’d not noted before that the evolved animal is like a software platform.

One of the curious facts about software platforms is that they aren’t good for anything. You have to pile an application on to the platform before it solves real problem. That is a useful right first approximation. Of course platforms are good for something, they are good for solving some space of problems. They allow you to build things.

There is a gap between a platform and a problem solution. In platform system design, where we don’t solve problems we just design more platforms, we think of these as layers.

For example the end-to-end principle suggests that the layers should be thin, so that the lower layers are windowed down to a kernel of necessary function and no more. In business theory where platforms go by other names like toolkits, standards, rule sets, and are observed in numerous guises such as major commodities on supply chain, we know that a platform creates an options space of further commercial activity. There is always a lot of competitive to and fro about who gets to capture that value. Suppliers often covet value created down stream from them in the supply chain. That’s no different than how platform vendors often fold high value innovations back into their platform offering. A move that is contrary to the end-to-end principle but is quite rational in a commercial mindset. When we complain about a supplier, say Microsoft, overreaching, say by bundling the web browser with the operating system, we call that monopoly; but as the quote above suggests it’s a kind of tyranny.

Reading and thinking about “Breakdown of Will” has been triggering some very surprising connections to all that. Animals are wired to manage their attention in a way that is at odds what we believe to be the optimal way to manage the attention of a rational man. There is a gap between the platform, i.e. the animal, and the problem to be solved, i.e. to be a rational man. It is into this gap that we humans pour our clever rationalizing schemes. Applications on the platform.

So that was my brainstorm. What triggered it was some stuff at the beginning of a book from the anthropology library about trying to explain religion. The introduction was working it’s way through the necessary dross and was talking about Darwinian explanations for religion. My reaction was “The platform can only tell you so much about the applications that run on it.” Darwinian ideas are a major supplier in the explaination of animal systems, but there is a tendency for people to let these ideas overreach their natural scope. You see a similar overreaching by the ideas that come of economics. At this point in my thinking about the ideas in “Breakdown of Will” I’m more inclined to put religion in the application layer as part of our struggle to create useful solutions atop the worse is better legacy platform.

What caught me was two things. This fun cheerful paper on “Self Experimentation” by the diet’s inventor Seth Roberts. What really did me in though was more than 30 years ago when I first became interested in cults I read my way through some marvellously silly books written by “Jane Roberts.” Jane’s gig was channeling, she would channel a dude name Seth. It’s a great exemplar of the art of speaking like a mystic; you know stuff like: “transforming invisible atoms into the dazzling theater of the world.” One side effect is that whenever I hear the name Seth I tend to get a foxy smile, and this time the name Roberts too!

I might not even have read the paper on self experimentation if it hadn’t been authored for inclusion in a book on behaviorism. I’m a huge fan of practical behaviorism; and I often recommend Jane Pryor’s book “Don’t shoot the Dog.” It’s delightful and a far better thing to read than this new diet book.

The diet turns about to be behaviorist at its core. Animals all (really all of them apparently) are very good at learning causal chains of behaviors; most of which end in food. The classic version of this is Pavloff’s dogs who he noticed would salivate when he rang the dinner bell; rather than when the food showed up. Animal trainers can do amazing things with these causal chains getting animals to walk around on two feet, jump through flaming rings, roll over, etc. etc. all just for a treat. The behaviorists have written libraries full of papers about the fine tuning of these causal chains, how to strengthen them, weaken them, extend them, etc. etc.

So the trick at the heart of Robert’s scheme is to weaken the causal chain between taste and calories. Consider the animal that has built a link between a bell and dinner. If that animal wants calories it craves the bell; it’s weird but true. Now of course a bell isn’t calories so we can weaken that link in two simple ways. We could randomly ring the bell so the animal abandons it’s illusion that these two things are linked. Plan B is we could stop ringing the bell before meals. Either will work just fine; though as the behaviorist research shows these links can be surprisingly robust if they have been trained up just right.

Robert discovered that both tricks appear to work. That he could reduce the body’s craving for food (aka taste) by either means. He could providing a lot of random tastes so it wouldn’t build a strong link between them. He even found articles in the literature of experiments where animals whose food was flavored somewhat at random – they stopped eating so much. He could also break the linkage by providing calories with zero flavor.    In both cases is the outcome is a weakened causal chain between taste and calories; which in turn leads to reduced craving for food.

The theory is somewhat more complex than I’m making it here. You’d have no trouble finding a few dozen explainations if you poke around in the web. But for me I was particularly taken to see a diet based on such an extremely simple confident application of behaviorism.