Businomics Blog

Financial Regulation: Paulson Misses the Trial and Error Economy

This new regulatory proposal from Treasury Secretary Paulson reminds me of the Patriot Act.  The department has a laundry list of changes it wants.  Then there's a crisis.  At which point the department pulls out its laundry list and says, "This is the solution to our crisis."  In reality, it's just the plan that they had on the shelf for a year.  If it really would have prevented the crisis, then the Treasury Secretary should be fired for failing to push harder to get that plan passed earlier.  But of course, that's not the case.

The Wall Street Journal editorial has the key insight into financial regulations:  the market is doing its job.  People say to me, don't we need some regulation to prevent lending to people who cannot afford the loans?  Here's what happened.

1. Lenders experimented with loans to people of lesser credit quality.
2. It worked in the initial stages, so the experiments were expanded into programs.
3. The large-scale programs failed miserably, causing billions of dollars of losses to the people who invested in the loan program.

Now, do we need laws to keep these investors from lending money to subprime borrowers?  I don't think so.  It will be years before those investors will touch that stove again.

If investors won't participate in large subprime lending schemes again, what would be wrong with some regulations?  After all, a regulation that prohibited me from hitting myself in the head with a hammer would not really crimp my recreational choices.  Here's the rub:  entrepreneurs need to experiment.  Some folks who are subprime may really, truly, be good credits.  If we regulate that loans can only be made to people with good credit, then lenders cannot experiment to find out if maybe they should change their lending guidelines.

A key element of our trial and error economy is that good business practices cannot be fixed in stone.  Companies experiment, try tightening a little, try easing a little, look for under-served markets.  For example, take "stated income" loans, in which the lender took the borrowers stated income at face value, without verifying the facts.  Sounds pretty stupid in retrospect.  So let's make a law that income has to be verified.  However, you've just redlined those folks with income that is inherently unverifiable: strippers, street performers, flea market vendors, etc.  We should allow lenders to test the waters with these people.

(If you are not a regular reader of the blog, take a look at some of the articles in the "trial and error economy" theme.)

Insurance regulation?  Part of the plan is a new federal insurance regulator.  What's the problem here?  My car insurer, home insurer, life insurer, are now regulated by my state.  I have not heard of any crisis, scandal or problem here.  This is another case of a "crisis" being used to justify totally unrelated regulatory changes.

Secretary Paulson says the plan will "reestablish the federal government's role in regulating the insurance industry by reclaiming a portion of its delegation of insurance regulation to the states..."  Whoa.  I'm not a lawyer, but I'm stunned that the Treasury Secretary thinks this way.  The constitution does not grant to the federal government authority to regulate insurers, aside from the federal government's authority over interstate commerce.  There has been no delegation of power from the federal government to the states, because the federal government never had it.

Incentive Systems: Markets, Medicare, and Your Business

"Markets and Medicare" is a great article in today's Wall Street Journal by my good friend John Goodman, president of the National Center for Policy Analysis.  The article has value beyond the Medicare debate, but let's first key in on those issues.  Dr. Goodman explains how medical costs can be lowered and health outcomes improved by freeing doctors, freeing patients, and freeing entrepreneurs.  Our rigid compensation system often will not pay for the most cost-effective treatment methods, so we get locked into expensive and less effective treatments.  Not good, of course.

If you're not interested in Medicare, here's how you should read the article:  your current relationships with employees, vendors, joint venture partners, and customers may also be too rigid.  If someone in the network of relationships around your company has a new idea, will the compensation structure allow that idea to flourish?  Do cost savings techniques get rewarded, especially if they increase customer satisfaction?  It may be time to take a fresh eye to your contractual arrangements.

The essence of economic progress is the Trial and Error economy (described in a series of articles listed here, my favorite of which is this one.)  Successful companies set up systems to encourage lots of little experiments in cost savings and customer satisfaction.  The experiments measure results, so that the winners are pursued, the failures are dropped, and progress is continuous.

Nike Strategy: The Trial and Error Economy

Nike has implemented a major change in strategy, according to an interesting article in Forbes Magazine.

The company had previously advertised with a mass market focus.  Its "swoosh" logo was everywhere, as was it's great motto, "Just do it."

The new approach focuses on specific sports markets: running, golf, soccer, etc.  Rather than having a footwear division and an apparel division, the company is now organized around specific sports and activities.  Is this another ho-hum moment in corporate strategy?  I've seen companies centralize to reduce costs, then decentralize to reduce costs, then re-centralize to reduce costs, each time taking a one-time special expense to pay for the reorganization.  I've become skeptical of corporate changes, but I like this one.

When the marketing focus is mass market, it's hard to say what's working.  If you are selling more Air Jordans, is it because of great design or the ubiquitous advertising?  Nobody knows.  And thus, nobody cares.  That is, without specific data, there is no specific accountability.

By working at the sport/activity level, the segment manager becomes responsible.  There's no passing the buck to the advertising manager; the segment manager has her own subordinate responsible for segment advertising.  If the skateboard segment is doing poorly, the VP of skates cannot blame the central design department, but only his own senior designer.

Trial and error does not work well at the amorphous global corporate level, which means the key method of improving products, production, and thus profits, is lost.  However, trial and error works wonderfully in a single business line, even if the corporate honchos don't understand exactly what they are doing.

Good work, Nike.

Starbucks Strategy vs. McDonald's: 10 Possible Competitive Responses

I've used Starbucks business strategy as a way to discuss the Trial and Error Economy (here and here).  I'm not, by any means, the world's greatest expert on the company, but it provides a great vehicle for teaching about corporate strategy.  Now, according to the Wall Street Journal, McDonald's will sell premium coffee drinks made by baristas at most of their 14,000 stores.

How should Starbucks react to the McDonald's threat?  Here are some ways:

1.  Do nothing.  Best implemented with one's nose high in the air, saying that Starbucks customers would never buy coffee at McDonald's.  Would work very well for three to six months.  Ignores the reality that Starbucks' recent growth has come not from Volvo-driving college grads, but from lower-income, less educated people than they originally served.  These are people comfortable at McDonald's.
2.  Cut prices.  This is the time-honored method of competition.  As an economist, I love price cutting.  As a business consultant, I almost always advise clients to avoid a price war.  In the case of Starbucks, I'd ask the company, "Who do you think has the lower cost structure?"  Not only should we look at labor costs, but consider this: at 9:00 am, McDonald's has lots of excess capacity.  Serving an additional customer is very cheap.  At the same time, Starbucks' chairs are full and there's a line at the order counter.  Serving additional customers means real estate expansion and hiring more staff.
3.  Differentiate the product.  In the classic form, the existing customer begins to differentiate, highlighting their product superiority.  Of course, Starbucks is already selling a product that it has successfully differentiated.  The practical way to do that now, in the face of McDonald's, is a "nobody makes a latte like Starbucks" campaign (using a catchier slogan that I just suggested, but pushing that theme.)
4.  Move upmarket.  In conjunction with more product differentiation, maybe Starbucks should raise its prices, ceding the price sensitive customers to Mickey D, but pulling more profit from the loyal customers.
5.  Lock up the resources needed to make the product. Starbucks is not going to corner the market on coffee, but in many cities they have cornered the market on corners.  That is, they have leased the top spots for urban coffee locations.  McDonald's, though, is known for great real estate.  (Not great buildings, but great locations.)  They are not optimized for the morning crowd, though.  Starbucks, for example, favors locations that are on the right hand side of the street for inbound morning commuters.  In rush-hour traffic, customers don't like to make left hand turns.  I'll bet that McDonald's has ignored this.  However, using existing resources instead of building new locations is a huge cost saver.  Bottom line: I don't think Starbucks can lock up a critical resource.
6.  Build customer loyalty.  Airlines, beginning with Western Airlines but quickly followed by American, built frequent flier programs.  Now a traveler has a big incentive to travel on one airline as much as possible.  But the potential rewards from buying all my coffee at one chain rather than another?  OK, Starbucks has their prepaid card which is really convenient; I use mine a lot.  They could add a rewards element to it; like prepay $20 and get $21 worth of sales.  But if I wanted coffee and I saw a McDonald's, with no Starbucks in sight, I would not keep driving just because I had some incentive card.
7.  Go head to head.  Imagine Starbucks looking for a location across the street or next door to every McDonald's.  Some of them might not be profitable locations, but it's a way to try to get McDonald's to drop the idea of specialty coffee.  Drive them out of that line of business.  I'm guessing that Starbucks already has plenty of locations near the competition,a but it would certainly be expensive to target all, or most, of the McDonald's locations.
8.  Merge.  One of them acquires the other.  They announce the Wall Street great synergies and economies of scale.   It would be really hard to make this work, but it's tried fairly often in other industries. Investment advice: sell short the merged entity.
9. Cut costs to maintain profits while sales are falling.  This is a pretty stupid plan, but common.  In Starbucks case, either the quality of the coffee would fall, or the best employees would leave.  The company would end up a shell of its former self.
10.  Ask for government regulation.  Probably won't happen here, but it's a standard response by plenty of corporations facing competition.  They might propose licensing baristas for the protection of customers, enforcement of antitrust laws against the competitor, product safety rules that would be hard for McD's to comply with (no meat preparation in the same location that coffee is prepared, to protect against mad cow).

There's the list of choices.  (Readers, feel free to suggest additional choices in Comments.)  How should Starbucks choose?  I recommend that they narrow the list to those that could be tried in a single market, rather than company-wide.  That would be 1 through 7.  (Maybe number 9, but that has really big potential impacts on company-wide reputation.)  Then Starbucks should its manager of each metropolitan area what they want to do.  Monitor sales data carefully.  Watch McDonald's, perhaps by hiring observers to count cars in the parking lots and going through the drive throughs.  Evaluate which response works best.  And of course, let metro area managers try other strategies I haven't thought of.

By trial and error, Starbucks should be able to identify the most successful response to the McDonald's threat.  The same process can be used in most cases of competition.

Starbucks Strategy: Trial and Error at the Beginning

Starbucks began with a willingness to experiment, according to Taylor Clark's new book, Starbucked: A Double Tall Tale of Caffeine, Commerce, and Culture.  The book begins with a story from when the company had only eleven stores.  The Vancouver BC store was at capacity, and another location in the city was hard to find.  Howard Schultz saw a failing restaurant across the intersection from the Starbucks location and rented it--not to replace the existing store, but in addition to it.  So the bit about Starbucks opening a store across the street from another store dates back to the beginnings.

They weren't sure it would work, but they tried it and monitored the results.  Actually, the monitoring did not require a sophisticated statistical analysis: both the old and the new store did a great business.

We learn from this story the value of the experiment.  It sounded stupid to me.  No amount of head office analysis would have justified the new location.  Some things you have to learn by doing.

What will you get if you try this approach?  A lot of failures.  Later in Clark's book there are plenty of stories of early plans that needed adjustments.  I'm certainly not endorsing bet-the-company risks every day.  But the businesses that succeed best are continually testing the waters with new ideas, new products, and new markets.  Even if the new market is just across the street.

(Want to read that story?  Go to Amazon's Starbucked page, click the "Search Inside" page, and search for "Experiment."  Look for the entry in Front Matter (where the introduction is).  It's one of several items; what you're looking for is the book's Introduction.  That will give you the first few pages.  The first real chapter of the book is available on the publisher's web site.)

Firefox Gets the Trial and Error Economy

The folks at Mozilla who put out the Firefox browser (which is great) understand the Trial and Error economy.  They recently wrote up a description of an experiment they did in marketing their download web site. (Hat tip to Steve Levitt at Freakonomics.)

Here's the test: if you go to a search engine to download the Firefox browser, you get the official website as the number one result from most of the popular search engines.  So should they pay Google or Yahoo for a pay-per-click ad on those search results when they are already number one?  It turns out that they get a little more traffic when they do.

One lesson from this, though, is that the majority of the traffic that comes from pay-per-click advertising would have come anyway; a majority but not all of the traffic.  So your cost per customer is not really what you are paying per click; it's your total costs of the clicks divided by the customers who would not have come to you without the pay per click.

In one experiment, it appears that about seven out of every ten people who clicked the paid advertisement would have clicked the regular search result (called "organic search") if the paid ad had not been present.  So all of your click payments only brought you those three out of ten who otherwise would not have come to you.

It takes a lot of work to get a good experiment, and you often have to keep running more experiments to get the results fine tuned.  But you can learn far more, and boost your profit tremendously, by regularly experimenting in your business.

eBay's Culture of Analytics: We All Need to Get on Board

Fast Company's latest article about eBay describes how the new chief technology officer created a "culture of analytics."  The old approach had been to set up a new feature or interface and see if it works.  If not, roll it back, returning to the old way.

The new approach was started by Matt Carey, the CTO.  eBay would show a new feature to a sample of users, about one to two percent of the total users.  They didn't know they were guinea pigs.  But their behavior on the site was tracked, in terms of how long they stayed on the site, how many pages they visited, how much business they did, etc.  Now one of the top developers says: "In a Darwinian sense, to be a survivor something has to keep producing."

The remarkable thing about this story is . . . that it's remarkable.  This should be cut and dried, basic business these days.  The data needed to evaluate an experiment at an on-line business are all available, if you just set up the systems to collect and analyze it.  Let's get on with it, guys.

(By the way, I've been watching what you click on to get to my page, and I've reacted.  Traffic is up.  Thanks.)

Business Goals vs. Strategic Intuition

Read the first two paragraphs of this Wall Street Journal book review:

Set big goals. Do whatever it takes to reach them. These muscular sentences form the core of commencement addresses, business-advice books, political movements and even the United Nations approach to global poverty. In "Strategic Intuition," a concise and entertaining treatise on human achievement, William Duggan says that such pronouncements are not only banal but wrong.

Mr. Duggan, who teaches strategy at Columbia Business School, argues that the commonplace formula has it backward. Instead of setting goals first, he says, it is better to watch for opportunities with large payoffs at low costs and only then set your goals. That is what innovators throughout history have done, as Mr. Duggan shows in a deliriously fast-paced tour of history.

(The full review is available here; subscription required.)

This message strikes home in two ways.  First, I've been blogging about the trial and error economy, in which companies adjust their actions as they get feedback on what's working and what's not.  That seems to be the essence of Duggan's book, which I have not yet read. (A couple of my key posts about the Trial and Error economy are here and here.

The second way this strikes home is that I've been working on a new business venture.  I'll post more about it here as it gets started, but it's a web-based information service.  As we talk to angel investors, they want to see our plans and goals.  Fine, I understand that.  But I have a forecast that three years into the project, at least half of our content will be driven by customer reaction rather than our own plans.

You could well ask if we should be starting a business if we don't know what to provide our customers.  But as we look around, hardly anybody is really doing a good job at monitoring customers to see what they want.  Here's an example:  suppose a supermarket manager noticed that customers were spending a lot of time in one aisle looking for something, and that half of those customers left the aisle without having placed anything in their shopping carts.  A smart manager would figure out that they are looking for something that isn't there.  Even the ones who buy something from that aisle may have been settling for second best; their time in the aisle indicates that they were looking for something rather than just grabbing their old-time favorite.

It's hard for the old-time supermarket manager to spot this behavior, identify the specific section of the aisle where this is happening, and bring in additional offerings to tempt the shoppers.  But it's pretty easy to do that on the web.  The biggest surprise is how little it's being done right now.

For the business leader or investor, I recommend: think less about plans and goals, and think more about how to identify opportunities and quickly exploit them.

Gum Marketing We Can All Chew On

Ben Pratt over at Market-Based Management Institute has a great post on his blog about selling chewing gum in Poland. Here's an excerpt:

He said that several years ago the Polish division of Wrigley ... had somewhere on the order of 70% market share of chewing gum sales in that country, almost none of it was the sugar-free variety.

Part of their sales and operations planning process was to document assumptions and revisit them at each planning cycle.  One of the assumptions they documented was that as people grew older their preference would shift from sugar-based products to sugar-free products, and demographics suggested that the population was aging.  At some point in the future they decided they'd have to change their mix of sugar / sugar-free products, but that time would be many years in the future.

Charles Koch's Science of Success quotes George Will as saying, "the future has a way of showing up unannounced."  And it seems to show up a lot faster than we expect.

Wrigley began seeing small but meaningful deviations to their expected sugar-free gum sales numbers.  Because they were revisiting their assumptions each month, they got curious and began testing whether the sugar-to-sugar-free preference shift was really as far out as it they'd assumed.  The result of their analysis caused them to recognize an opportunity: people in Poland were looking for sugar-free products, but not many companies were supplying them.

This information caused them to re-think their strategy and they immediately shifted their product mix almost entirely to sugar-free gum.  They captured 90%-plus of the market.

I'm guessing that without a formal sales forecast as a benchmark, the company might not have caught the shift in trend.  A minor product grows a little faster than expected; ho hum.  But the trial-and-error economy demands that you watch for the little things that are growing faster than expected.  Then be ready to nurture them into really big things.

How Quants Go Wrong

Here's a great quote from a recent Wall Street Journal article:

"Events that models only predicted would happen once in 10,000 years happened every day for three days."  Matthew Rothman, Ph.D., Lehman Brother Holdings

I hate to take pot shots at analytical guys--I'm one myself.  After all, you don't go to graduate school in economics just to pick up chicks.  (Well, that's one reason among many.)

Let's ask, how could they be so wrong?  Simple.  They were not remembering their assumptions.

Remember the normal curve (also called the bell shaped curve)?  It looks something like this (courtesy of Wikipedia):

Suppose we have a bunch of data, like on daily prices.  (More likely, on the change in the logarithm of the price.)  We can look at the data and see if it fits the normal curve.  But wait.  We don't have many observations out at the tails of the curve.  Like things that should occur once every 10,000 years.  We may not have any observations, or we may have one.  In any event, we can't really tell if the tails of the distribution fit the normal curve or not.  So we look at the great mass of data in the middle, say that it looks like it fits the normal curve, then assume that the tails of the normal curve also fit the data.

But they don't.  We've known for years that financial transactions are "leptokurtic," which is a great way of saying they have fat tails.  In other words, there are more extreme values than the normal curve would predict.  I assume that Dr. Rothman knew that, but what does one do with the knowledge?  Assume some other distribution, similar to the normal curve, but with fat tails.  Unfortunately, it's very hard to know just how fat the tails really are.  We can collect many, many years of data and still not know very much about the far tails of the distribution.

Great example:  In 1997, my local newspaper reported that the Tualatin River (near my home) had surpassed the hundred year flood mark, but "it wasn't as high as last year's flood."

What's the moral of this story?  We have to be very cautious about extreme events.  We don't really have a clue how frequent they are.  That's true of financial markets, but also true of physical sciences.  If an engineer tells you the boiler has one chance in 10,000 of blowing up, I'll bet that the engineer does not really know the true odds--he's just extrapolating on the assumption that the distribution is nice and neat.

What to do about it?  Don't place too much confidence in hedging strategies.  Instead, plan for bad events, and be sure that you know how to react when they happen.