Bryan Caplan  

Come One, Come All: The Mystery of Business Tolerance

Kipper- Und Wipperzeit Update,... How Capital Regulation Creates...
Firms are usually picky about who they hire.  But when they spot potential customers, their standard slogan is "Come one, come all."  Sure, there are some exceptions.  A few restaurants still have dress codes, and some rental car companies won't rent to under-25s.  But by and large, firms welcome paying customers of all shapes and sizes.

It's tempting to shrug and say, "What's the puzzle?"  Firms are picky when they hire because workers are not created equal.  But they're promiscuous when they sell because money's money.  Right?

Wrong.  A customer isn't just money.  He's a package of money and personal issues.  If you're running a restaurant, some of your customers complain and dawdle, while others take what they get, then eat and run.  If you're running a hotel, some customers leave their rooms in pristine shape, while others treat their room like a pig sty.  And if you let customers return merchandise, some of them never exercise their option, while others return more than they keep.  If businesses charge every customer the same list price, you'd expect them to prize some customers and dread the rest - not welcome all comers with open arms.

You might blame discrimination laws, but don't be hasty.  Only a few kinds of discrimination are actually illegal.  And in any case, firms have plenty of tools to discourage undesirable customers without breaking the law.  Strict dress codes are only the beginning.  You could simply jack up your prices, then target discounts to the right zip codes.  As the upscale mall on The Simpsons brags, "Our prices discriminate because we can't."

So what's the real story behind firms' "Come one, come all" ethos?  Some leading hypotheses:

1. Despite some vivid examples, customers rarely predictably vary much in profitability.

2. Businesses have already equalized profit margins with frequent flyer programs, selective coupons, and other preferred customer programs.  Good customers get good deals, bad customers bad full price, transforming everyone into a cash cow.

3. In less tolerant times, businesses were more selective.  But most moderns would be embarrassed to shop somewhere snobbish enough to turn away paying customers.  The upshot is that good customers indirectly subsidized bad customers to avoid feeling elitist.

4. Deviant behavior by both workers and customers has increasing marginal costs.  If 5% of your workers or customers are troublesome, it's no big deal; but 10% is more than twice and bad, and 20% is more than four times as bad.  At least for smaller businesses, then, hiring a single difficult worker is a big problem, but a few difficult customers are a light burden to bear.  An implausible prediction is that big firms will be more willing to give weird applicants a chance.

Which stories make the most sense to you?  Got a better explanation?  Please show your work.

Comments and Sharing

COMMENTS (33 to date)
Brian Moore writes:

Isn't there a whole line of business thought that says that certain customers aren't actually worth it? I'm thinking of the Best Buy (???, maybe circuit city) policy where they tried to "buy out" customers that were actually costing them money. Or perhaps the American Express policy of paying certain high-cost people to stop being customers (again, my vague memory only)

Tim Ferriss (the 4 hr workweek guy, take that for what it's worth) applies the 80/20 rule to customers -- a small minority can take up lots of your time, so if they decide not to do business with you, then you're better off.

I do agree that almost every business wants to appear as if it has a "come one come all" policy, and even when selling "exclusive luxury" or conspicuous consumption items, the theme is "yes, you too are worthy of our product!" rather than "please don't."

Another point might be that we've pushed most transaction costs so low that unless a customer is especially costly, most sales are good sales -- even loss leaders that drive other types of sales.

david writes:

Prices don't clear perfectly for assorted market-power reasons, and at equilibrium businesses prefer continue selling stuff at the prevailing price and employees prefer to continue working more at the prevailing wage.

[employers changed to employees by request--Econlib Ed.]

Bill Conerly writes:

When I get together for refreshments with my fellow consultants, we all have stories of customers that we fired, or declined to do business with in the first place. Typical line: "I'm not the best person to help you on this project."

There was a business best-seller a few years ago, Angel Customers and Demon Customers,which made the point that some customers were not profitable.

The best companies set prices and terms such that unprofitable customers choose to do business somewhere else. If the companies are really good at it, the customers never realized that they are being pushed out the door.

wintercow20 writes:

I like your list, but since you asked for other possible ideas ... perhaps firms/workers use diverse groups of customers as a learning tool. So, the marginal revenues from selling to an "undesirable" customer may in the long run exceed the price.

I am not sure how convinced I am by this, but I like the exercise.

joeftansey writes:

Disassociation is a lot easier. You just tell the customer to leave. A troublesome employee usually has a contract, and you also need to go through the trouble of finding a replacement. This is complicated by training costs and the like.

Chris writes:

How about a fifth option - prices are a way to pre-select consumers. Restaurants that charge extra for atmosphere will pre-screen for customers that appreciate atmosphere. Customers that appreciate atmosphere will be less likely to do something to destroy that atmosphere.

Seth writes:

I like your stories. I have one to add.

Daft managers. See the book "Manage for profit, not for market share". I've worked with more than my fair share of managers and board members who strongly believe that "Wall St." values client growth, even if they have to achieve that growth by reducing revenue and earnings (some even believe stock analysts control the stock price) or just simply view business as if it's a baseball game where the score is kept with client counts. I'm amazed these guys continue to get jobs. But, I also have a theory that true talent is rare and concentrated in organizations that look for it, the rest are faking it to make it. Even the NFL, with a controlled size, has a hard time finding good enough talent to round out all of their squads (and in many years more than 10 or 15).

Something I see from folks whose incentives are tied more directly to their bottom line (as in their bottom line buys their groceries tonight) is that they have a much better understanding of the value prop they offer, who wants and values that prop and they focus on delivering it for those customers. They don't feel like the girl left out at prom if a client turns them down.

Brandon Berg writes:

I'm not sure I buy the premise. I think that insofar as firms are able to identify unprofitable customers, they try to discourage them, and in extreme cases even flat-out refuse to serve them or ban them from the premises.

This is especially true with insurance companies, who've made a science out of identifying unprofitable customers. This suggests that insofar as other types of firms don't refuse to do business with unprofitable customers, it's because they can't identify them reliably enough to make it profitable to practice statistical discrimination.

Remember the auto insurance company---Progressive, I think---that used to offer to give you quotes from three competitors? I've heard a too-good-to-check explanation for this: For most customers, they'd have some competitors whose quotes were higher and some whose quotes were lower. When they expected a customer to be profitable, they'd show only quotes higher than theirs. When they expected a customer to be a money-loser, they'd show a competitor with a lower quote in order to encourage the customer to take his business elsewhere.

This is brilliant if true: They were able to engage in statistical discrimination while creating the appearance of being honest enough to tell customers about competitors with better prices.

Evan writes:

I think 1 sounds the most likely. We all have spectacular examples of customer idiocy. However, for every one of those, there's a zillion customer we don't remember because they were so blandly normal. It's classic availability bias. The other commenters have mentioned have mentioned training to deal with difficult customers, but most of those seem to deal with how to get rid of a bad customer you've already met, not how to stop them from coming in the first place.

3 is also convincing. Our society is more egalitarian than it used to be, we tend to frown on overt status signalling. I know that whenever I hear someone scornfully and condescendingly putting down this or that group of people, I immediately sympathize with those people (even if I didn't before) and conclude that person is far more unpleasant than the people they're denigrating. I think any business that blatantly tried to stop "person type X" from doing business with them would make a lot of good customers avoid them.

Norman writes:

"An implausible prediction is that big firms will be more willing to give weird applicants a chance."

But given that many firms don't figure out which workers are weird until after the hiring process is over, a much more plausible prediction is that big firms will be less likely to fire weird workers after they identify the weird ones.

Tracy W writes:

Like Bill says, in my experience of consulting we've turned down clients occasionally because their past behaviour has made contracts not profitable (and in one vivid case, the project would have consisted off a lot of really tedious work, which they would then have picked over with a fine-toothed comb.)

And I did some programming work for a credit card company once where they were working on ways of discouraging customers who take full advantage of their 0% introductory offers deals.

I'd go with 2.

Toni writes:

How about the fact, that with a lot of companies having no efficient way of remembering bad customers, and with discrimination of bad customers really difficult, it's just not practical?

I mean, when a guy walks in to your restaurant, how exactly do you see if he's a bad customer?

You can put potential employees through a lot more due diligence than you can your customers.

David Jinkins writes:

While I was in high school, I worked a summer as a hotel receptionist at a budget hotel in a small Wisconsin town. When the fair came through town, my boss (the owner) told me not to rent to carnies. She said that they rent one room for everyone to shower in.

Matt writes:

I think it has more to do with the fixed costs of assessing customers vs employees. An average employee is going to have a salary of around 40k-50k per year, so it makes sense to spend some time and money to make sure you have the right employee. Very few interactions with customers are of that magnitude though, so those fixed costs just eat into profit margin. And with businesses that do do sales at that level, you start to see standards emerge. A 20 year old in a track suit isn't likely to get much attention from sales people at a luxury auto dealership.

I think standards are also more likely to emerge where repeat, face to face business is frequent (see country clubs and 5 star restaurants) than when it's absent (eg home purchases and vacation resorts).

Daublin writes:

I don't agree with the premise, either. If you trash a restaurant, and they recognize you, you are not welcome when you come back. You get the benefit of the doubt on your first visit, but to the extent they can, restaurant owners will weed out the baddies over time.

I always thought the saying was more of a psychological motivator. They need to remind themselves that even for people they'd avoid a social relationship with, it is worth the effort to have a business relationship.

Scott Gustafson writes:

Bryan needs to take a class in Services Marketing.

In marketing you worry about the 4 P's - Product, Pricing, Placement and Promotion. Any reasonable marketing class will cover those areas. The trick is that those things work for physical products. For Services you've got 3 additional P's - Process, Physical Evidence and People. It's the People part that answers his question.

Services are produced and consumed simultaneously. (If you ever figure out how to store a service, let me know. We're going to make a lot of money.) Services are often face to face. You're working directly with the customer and you have to adjust your service delivery based on your real time interaction with that customer.

In many cases you are dealing with several customers at the same time. Customers can interact and that will affect the delivery of services to each customer.

I teach economics at Mesa Community College. I deliver educational services. I have up to 35 students in each of my six classes. As any student will tell you, one disruptive student will ruin the class for everyone else. Given that problem, you have some options.

You can let them ruin the class for everyone and decrease the value of your service. This negatively impacts the future profitability of your classes.

You can fire the disruptive customer (drop them from your class) and carry on with the remainder. This decreases profitability but maintains the value of the delivered service for others.

You can deal as best you can with the disruption and try to minimize the impact on other students. This maintains the income stream from the problem student while minimizing the detrimental impact of others.

At the Community College we take all comers so that rules out the option used by other institutions - Admissions selects who shows up. This helps ensure a quality experience for all involved.

Every service delivery organization makes these choices. In rare cases, they do actually fire customers. They do that because certain customers are not only not profitable, but reduce profitability with their other customers.

With that preface, let me answer Bryan's 4 questions directly.

1. Service delivery business owners can very quickly assess whether a potential customer is profitable or not. If they can't, they won't be in business very long. As a test, ask any server in a restaurant. They'll point out the big tippers and the lousy tippers as they walk in the door.

2. Yes, good customers are offered better deals than bad customers. This is especially true in professional services where it is easy to price discriminate.

3. Tolerate times or not, service businesses are selective. The selection methods may be subtle - not marketing to a particular potential client, or overt - suggesting that a client go elsewhere next time or refusing to take a customer as a client.

4. Bad behavior on the part of service delivery personnel is easy to deal with in the private sector. They get fired. In the public sector, it is more problematic. Bad Behavior on the part of customers can also be dealt with. In most cases their price goes up. In other cases, they get fired (asked to go elsewhere or dropped as a client.)

If you would like examples, just talk to service delivery business owners, especially those that focus on professional services - doctors, lawyers, accountants, financial advisors... and education professionals.

Luke writes:

Drawing on what Bill, Tracy, Seth, and Scott have said, it seems like the real issue is a lack of local knowledge. In service industries where the decision to "fire" a customer can be made by front-line employees, it does occasionally happen. If management policy decrees that "the customer is always right", employees are forced to pretend that this is so.

Noah Yetter writes:

I'll pile on in rejecting the premise. Customers are turned away more often than you think.

(If you ever figure out how to store a service, let me know. We're going to make a lot of money.)

It's called software.

Lint writes:

I'm going to go with the first theory. Perhaps my experience working as a delivery driver for 6 years can shed some light on this.

Servers will rarely ask tables to leave, even if they know they'll get a bad tip from that particular table. You might be tempted to think that it's because other tables would notice the server being snobbish and would lower their tip accordingly, but I don't buy that. The reason that sounds unconvincing here is that they rarely ask tables to leave, even when there is only one table in the restaurant. The only time I've seen people be kicked out is if they were particularly rude or bothersome to other customers, and then it would be by management. So do they not throw people out because they fear management getting angry at lost revenue?

But what about delivery drivers? Work long enough and build a good repertoire with management, and you can get away with whatever you want, provided you're still being productive. Now something I've done, albeit VERY rarely and only with repeat customers, is instruct customers to order carryout to avoid paying a delivery charge. I obviously only instruct bad customers (i.e. bad tippers who live towards the edge of our delivery area) to do this, and usually this doesn't upset them, especially if I hide my anger over not being tipped well. This also works very well when frequent bad tippers call in to place their order. When a delivery driver familiar with that particular customer (or really the address, since that's how we remember things) is the one who answers the phone, we often point out how much they could save by picking it up themselves. CSRs, cooks, and managers rarely do this. But the key thing here is that I'm out of the view of management. When a server would throw out a table, management can clearly see that. When I convince a delivery order to switch to carryout, all management sees is a carryout order. Note that convincing them to do carryout is completely in my self interest since I get paid a lower hourly wage when I'm on the road than when I'm in the store.

Looking at this scenario, if there is reason to discriminate against customers, we should see it being done by employees who work for tips moreso than employers, who receive the same revenue regardless.

Chris Lemens writes:

I reject th epremise for some companies, but not others. The underlying issue is the cost of obtaining useful information about customers. If you are in a mass market (think big-box retailer), then it is very expensive to obtain useful information about your customers, because there are so many of them. And even if you get the information, acting on it may be difficult -- will you have the guy at the front door check ID as shoppers enter? If you are a consultancy with 10 customers at a time, where a lot of your business comes through referrals, then information is cheap. And it is easier, by comparison, to act on the information because the clients are not protected by anonymnity.


tribsantos writes:

I think that a lot of it has to do with iteration. You will have to deal with a worker for a relatively long time (people don't get fired before a month, usually, and if a worker stays for longer than that, he can make friends in the firm and his firing could harm morale). So you need to get that right.
A customer has usually only one interaction. If you spend too much time trying to identify who will be the guest that will shout with the waiters, you may lose many perfectly good customers.
I think also that, even though you expect more from your customers than just money, you don't expect a lot more, so customers are a lot more alike than workers. In a restaurant, you are supposed to dress well, treat the waiters nicely and don't talk too loud. The great majority of people that are willing to afford a meal at restaurants are able to do that.
If you think of markets for non-iterated workers non-qualified tasks, I think that the firms hiring behave pretty much the same way as firms attracting customers. Think of movie extras.
Conversely, firms that are looking for long relations where you expect some kind of more specific behavior from customers scrutinize more (law firms, psychologists...)

Eileen writes:

Don't rule out that people inside the firm have competing goals and incentives.

Sales and sales management are paid on things like revenue -many time this is regardless of profit. Not to mention that their jobs hang on reaching quotas every quarter and sales is a powerful department in most companies.

Management may not even calculate profitability per customer, so though a few people may understand that a certain customer is not worth the revenues they bring in, they fight an uphill battle.

By the way, in my experience, the government is the biggest "bad customer" based on resources required vs profit generated. Companies continue with them in hopes of getting bigger deals or just for the bragging rights of selling to them.

Adam writes:

Businesses go to great lengths to select customers in ways that don't cause legal problems. Brick and mortar businesses choose their locations. Saks and Nordstrom don't locate in a local strip mall. Subway's all too happy to be wherever anyone's likely to be hungry and have 10 bucks. Click businesses have detailed strategies for recruiting clicks that sell. Retail business use special mailings and discounts to recruit buyers. Kohl's is one familiar retailer that seems to have a highly refined mail-based strategies of personalized pricing and discounts.

drobviousso writes:

I've noticed, in my own work experience, that the larger the investment in a customer, the more likely the firm is to turn a bad customer away.

In fast food and big box retail, we rarely turned away a customer that wasn't a thief or an active problem for another customer.

In college and legal firms, where each customer is a huge investment before a profit starts showing up, we are much, much more selective.

Nooman writes:

Isn't it as simple as:

"the customer may be King, but I am your BOSS".

tribsantos writes:

Complementing my commentary, if one asks why is it so more common for employment to be permanent and consumer transactions transitory, I think that it is a core object of the theory of the firm.

Brian Clendinen writes:


Maybe in retail that is true, however, in construction with large project that is not true. In a bad economy rarely will we not bid a project we have a decent chance of winning, but we will price a lot higher profit margin to make up for the unknown cost of dealing with a bad customer.

On the other hand Seattle Airport had major issues in screwing over its contractors and has had all sort of law suits. Now no one in the industry wants to deal with them, if they do, I would estimate they are about 40% higher than any other Washington project would be.

So the method is we don’t want your business so we will give you a really high price. If you accept then we get paid a premium for the pain and suffer you cause us. When it comes to governments, people just price the pain into the picture.

Nathan Smith writes:

I disagree with those who reject the premise. Customers are sometimes rejected, but the pattern of "come one, come all" is very general.

I think all of Bryan's explanations are superficial, though they overlap somewhat with mine, which are as follows.

1. SPECIALIZED PRODUCTION, DIVERSIFIED CONSUMPTION. Start with two deep general facts: (a) there are fixed costs and increasing returns at the individual level in the production of just about anything, and (b) there is diminishing marginal utility from consumption of just about anything. Development consists largely of people becoming more and more specialized in what they make, but more and more diversified in what they consume. So there are systematic asymmetries between the production and consumption sides of markets.

2. MARKET COORDINATION. Think about transactions are coordinated. This is a problem neoclassical economics essentially ignores, but economists have to think about it when they're dealing with stock markets. In stock markets there are (to simplify a bit) two kinds of orders: (a) limit orders, and (b) market orders. A limit order has a fixed price, e.g., $20/share, and executes whenever the price is offered. A market order executes as soon as it comes in, at the best available price. Please note the asymmetry here. Every transaction (basically) is of a market order with a limit order. Two market orders can't transact, because they would have to coincide exactly in their time of arrival. (Also, what would be the price?) Two limit orders can't transact, because they would have to coincide exactly in price. (Also, *when* would the transaction occur?)

Now, generalize this to other markets. You still need a market order and a limit order. One transactor has to be waiting when the other takes the initiative. And inasmuch as producers are specialized, i.e., they're always doing the same thing, while consumers are diversified, i.e., constantly varying, it makes sense for the producers to do the waiting. Think of a person who writes a shopping list, with offer prices, on a large white sign, then sits out by the highway waiting for someone to come along and do business. Absurd, no? The producer makes limit orders, the consumer, market orders. That's why businesses say "come one, come all." They are the limit orders, just sitting there, letting customers take the initiative.

3. INCOMPLETE CONTRACTS. The other question, why do firms NOT say "come one, come all" to workers? Here we need to consider the nature of a firm, and in particular, Coase (1937)'s question of why there are "islands of planning in a market sea." Why hire a worker as an employee, rather than hire him on a contract-by-contract basis to specific jobs? Because you can't write a contract to cover every contingency. Also, you can't perfectly monitor the worker's effort. So you want to know about the worker's motivation and character, stuff that's hard to observe. Of course, incomplete contracts is to some extent a problem on the customer side, too, which is what motivated Bryan's original puzzle.

But please note that if there is NOT a problem with incomplete contracts, you don't need an EMPLOYEE at all. You can hire the same worker as a contractor. In that case, the boss becomes a customer. Imagine for a moment that all contracts could be complete, there would be no need for firms or employment contracts, and everything could be done on a piece rate basis. Even then (indeed even more so in this case) producers would adopt a "come one, come all" rule vis-a-vis customers, because they play the limit-order role in the market coordination process. Meanwhile, workers would turn into mini-firms, and they might even adopt a "come one, come all" rule vis-a-vis THEIR customers, i.e., bosses.

I think this is the basic reason why businesses serve all comers as customers, but hire selectively. Other reasons are of limited importance. And some exceptions aren't exceptions. A restaurant with a dress code is still saying "come one, come all," only the price includes both money and dressing up.

John Fast writes:

Why not simply ask some real business managers why they aren't more selective about their customers?

Bryan Willman writes:

The premise is false in many non-retail circumstances.

For example, there are a whole host of implicit and sometimes explicit "go away" messages from industrial suppliers.

Grainger is famous for "only selling to business" and will sometimes actually make people prove a business association. One assumes liability law or the like is the reason.

*Some* industrial suppliers will just hang up the phone on callers who are "too small" - not only will they not sell you 1 widget, they won't sell you anything until your volume reaches some level (like $100K to $1M a year.)

Others just tell you "really only interested in accounts bigger than ...."

Customer ride race teams (paying drivers) alternate between being desperate for money and very selective who they'll take as drivers (customers) - the reasons why aren't hard to grasp.

So, lots of businesses ARE very selective about who their customers are. Retail is probably sometimes a special case.

Miguel Madeira writes:

a) In an organziation where every worker has a specific function, a bad worker affects not only his own productivity, but also the productivity of their co-workers. In an extreme case, in a line-of-production technology all workers have to work at the speed of the less-productive worker.

In contrast, enve if a customer gove low profits (or even create a loss) this usually does not afect the profitability of the other clients.

b) The dimesnion of the staff is usually fixed; this mean that hiring a worker implies refusing to hire another potential worker; them you have to choose carefully what workers to hire

In contrast, most bussinesses does not work at their full maximum level of production - then, there is allways room for one more client, meaning that you don't have to choose between customers (even if you accept a bad costumer, you can also accept a good). Note that, when bussinesses really work at their maximum (for example - a crowded nigh club) there is indeed selection of customers.

Charles R. Williams writes:

It is one thing to say that customer x is not worth targeting. It is another thing to turn him away at the door. And it is another thing entirely to train your people to welcome customer y and be rude to customer x, who might just be customer y's cousin.

Businesses do target desirable customers and discourage undesirable ones but this is done with a smile.

John David Ward writes:

I think it's a principal-agent problem. The business owner wants to discriminate to maximize profit, but he or she probably won't interact directly with the customers. Some employee will. But the interests of the employee are different from the interests of the employer. The employee wants to minimize the amount of work he or she is doing. So the employer removes the employee's power to discriminate, except according to certain clearly laid out rules, to prevent it from being abused.

Comments for this entry have been closed
Return to top