Tuesday, July 31, 2007

And with that...

OK. That post was incredibly long. I need a vacation.

Actually, I figured I'd start blogging when noone was around to try and get the hang of it. Now I'm off to Cape Cod to eat fried food take a long-planned vacation. I doubt I'm going to attempt to post from my Blackberry, so have a good beginning of August and see you in a couple of weeks!

Win a Free iPod!

So which newspaper has the best business news? The New York Post, of course. In fact, the New York Post is the only New York newspaper that seems to have business news. Everybody else has analysis. Plus, they delight in covering the media, marketing and internet business in New York. No one else really does. Every morning I open to the business section of the Post with trepidation. I never know what I'm going to find. Anyhoo, just had to say that.

From their article today on ValueClick's missed earnings:

ValueClick's revenue from so-called lead generation business, which includes online sweepstakes, fell during the quarter after the Federal Trade Commission started investigating the practice in May, Chief Executive Officer TomVadnais said.

The FTC is investigating whether promises of free gifts violate U.S. laws.

I have no idea if this is what ValueClick said on their call. I didn't listen to it. But it's a perfect illustration of what is going on in the "so-called" lead generation business. (What's up with that "so-called"? Is Google a "so-called" cost per click business? Weird.)

Lead gen, if you're one of the thirteen people in the United States who hasn't heard my lead gen spiel yet, is the next step in the evolution of marketing from being push to being pull. I vaguely mentioned last week something about lower search costs changing the nature of marketing. One of the ways this is happening is that consumers are now paying more of the search cost and marketers are paying less.

Wait! It's not like that. It's like this: marketers pay the same average dollar amount per customer, but take less risk. The media get paid the same dollar amount by the marketer, take more risk, but make the consumer do some work. The consumer 'pays' more of the search cost by providing extra information about themselves in exchange for finding a product that better matches what they are looking for.

Am I being obtuse? Take Google as an example. Instead of running banner advertisements, they show cost-per-click ads. A marketer who buys a CPC ad on Google pays the same amount per sale as running a banner ad but takes less risk that his money will be wasted. An auto dealer could run a display ad at a $25 CPM and expect one out of every 20,000 people who see the ad to buy a car from him. Or, he could buy a click-through from Google for $10 and expect one out of every fifty people who click on it to buy a car from him. He takes less risk with the second option because he doesn't pay unless someone clicks and people who click are more likely to become customers. (In the real, inefficient, world right now, the dealer actually pays quite a bit less with the second option, and we'll get back to ValueClick with that fact later, I promise!)

The consumer provides information to Google by typing in "Honda dealer NJ new car", or something of the sort. The fact that the consumer is looking for a new car Honda dealer in New Jersey is valuable information and she is giving it to Google in exchange for information about these dealers. Google is, in turn, using this valuable information to put up CPC ads that point to New Jersey Honda dealers. By only getting paid if the consumer clicks through, Google is taking the risk that it will squander the consumer's valuable information and not get paid by an advertiser. (As an aside, it would be interesting to know how much Google is getting paid for taking that risk, consumer shouldering of search cost aside; does anyone know Google's overall effective CPM? I guess it shouldn't be hard to figure out. Knowing that would, if you believe in efficiency, put a dollar figure on the value of the consumer search.)

Lead gen is the next stage in the evolution of product-consumer matching from being a cost borne mainly by the marketer to being a cost borne mainly by the consumer. CPC is to CPM what Cost-per-Lead is to CPC: a further outsourcing of customer acquisition risk by marketers and a further increase in consumers trading information about themselves in order to find more appropriate products. This process is as inexorable as the move from CPM to CPC, from Doubleclick to Google. The next Google will be in lead generation.

But yesterday's news says it won't be ValueClick. The problem with outsourcing your marketing, or with outsourcing anything for that matter, is the familiar principal-agent problem. If you are paying someone to provide you with something that looks a certain way, they will provide you with that in the cheapest way possible, even if providing it cheaply strips all the value from it. For Google this looks like click-fraud. For lead gen it often looks like incentive marketing.

Ever see the ads for a free iPod or a free flat screen TV? (Maybe not so much anymore because of the above-mentioned FTC investigation.) You are promised a free iPod if you fill out a bunch of personal information. Next thing you know, you have a mortgage broker on the phone, trying to get you to refinance your mortgage. You've been sold as a mortgage lead.

You are a bad lead, obviously. Not because you aren't a real person or because you don't have a house to refinance (those would be pre-requisites for selling you as a mortgage lead) but because you have no intention of refinancing. A lead has to have both personal information and intentionality. If the person who filled out the lead form does not have the intention to purchase, then they are a bad lead. A company that entices low-intentionality people to fill out lead forms is selling poor quality leads. Lead buyers can figure out a given lead's quality on a lot of dimensions: is the person real, does the name match the address match the telephone number, if the lead becomes a customer then how much will that customer be worth, etc. But determining the intentionality of a given lead is very difficult.

That's why some lead generators sell low-intentionality leads: they are cheap to generate and indistinguishable from high-intentionality leads for a long time. An extension of Gresham's Law holds: bad leads drive out good. And because leads are fungible (a low-intentionality lead generator can sell his leads to another lead generator to be sold to a marketer) the industry has accrued many bad, or at least cynical, actors and the marketers can't weed them out.

Now, I'm not pointing fingers at anyone. I don't know ValueClick. But incentive marketing creates low intentionality, almost by definition. The mortgage broker calls and the prospect say "yeah, yeah, where's my iPod." One reason marketers pay less, on average, for a customer acquired through lead gen than through display advertising is that marketers don't trust lead generators to not screw them.

This is a problem the lead gen industry needs to solve before our Google can push the last Google into Doubleclick-like obscurity. Any takers?

Why Doesn't *My* Argumentative Ultrarationalism Impress Women?

I love

Monday, July 30, 2007

I Won't Even Comment on the Business Press

Matthew Yglesias over at the Atlantic has an incisive take on the annoying bias of journalists towards print media. Even notably rational tree-killing news sources like the Economist get all preachy about the internet when they're reminded that having somebody hand deliver the words their writers set down literally days before is just kind of silly.

Yglesias comments on Russell Baker's crotchety waving of the walking stick at the younguns in his review written in the NYRB:

"The Internet" can't replace The Los Angeles Times's congressional coverage, but the congressional coverage of The New York Times, The Wall Street Journal, and The Washington Post (plus the wire services, plus the non-print media) plus The Hill plus (if you're willing to pay) CongressDaily, CQ, and Roll Call most certainly can. What "the Internet" can do is make it very, very, very easy for a person in Los Angeles to access that kind of coverage.
In other words: you don't need a local paper for national and international news. You don't need an airplane full of reporters following Clinton around to know what she says in her latest rendition of her stump speech. You just need a couple. Yglesias goes on to point out that the existence of too many newspapers is good for one group: newspaper writers. If you have fewer newspapers, you'll have fewer reporters. Implicit in this is that you'll actually have better journalism: the surviving reporters may well be the best ones money can buy.

The unstated dark side: is the country ready for the rise of the celebrity reporter in the print media? The one nice thing about journalists getting paid far less than they could make in other fields of endeavour is that we know they're not in it for the money.

Thursday, July 26, 2007

The Virtue of Being Completely Insensitive

Rejection Sensitivity (or RS)--for those of you who don't read Psychology Today--has been on the rise of late. Parents are overprotecting and overpraising children. (Well, except me. As noted previously, I'm training mine to be professional negotiators by making nothing easy.)

Over at Free Exchange they're blogging about how the various Facebook Crush apps can reduce RS. They make the point, in their inimitable English way, that this will reduce the income inequality between men and women because men have historically had lower RS. After spending their formative years being rejected by women, men become rejection insensitive. According to the blogger, the average male has been rejected scores of time by the time they are 25. Yes, yes, I know, but I've spent time in England and will verify that Englishwomen are notably hard to charm. Or maybe just for those of us who read the Economist, I don't know.

Their point is that having a low RS is a benefit in the risk=return world of business. I have a different worry: every entrepreneur I've ever met has a vanishingly small rejection sensitivity. If these Facebook apps raise RS, then there will be no entrepreneurs, no startups, and our economy will be cooked.

This is a disaster in the making. What can we do to stop the Facebook Crush apps, before it's too late?

It's Just Like a Library, So Please Stop Talking to Me

I went to Barnes & Noble at lunch. I like to buy books, it makes me feel better. Not reading them, just buying them. They look so interesting. Someday I'll get around to reading them all.

Oh, who am I kidding?

At Barnes & Noble the past few weeks I've noticed that I can't stand still without someone asking me if I need any help. Not just the nearest B&N, but at the four or five B&Ns I've been to in the past months (yes, well, they're all different.) I assume this is a corporate strategic change: an emphasis on customer service! Yay!

No. At Home Depot I need customer service. I do not know what plumber's putty is, how to use it, how much I need and where they hid it. I need someone to help me. If I needed help buying a book, I would go on the internet. When I go to the store I want to find things I didn't know I thought were interesting. How can customer service people help me find something serendipitously? Why do smart companies keep trying to compete when they don't have to? Don't compete with Amazon, do something else, something you can do better.

And no, the B&N customer service person could not help me with my questions about plumber's putty.

Wednesday, July 25, 2007

Because We're Creatives, Not Bean Counters

Way back in 2002 (or 2003) The Economist wrote something along the lines of: the marketing department is the only place in the corporation that has not adopted a system to make their job more measurable and efficient. That this was true then was astounding. That it's true today is mind-blowing.

My friend Rob Leathern wrote about a recent frustrating media buying conversation. Why, he essentially asks, do I have to go all or nothing in a buy, hoping that it works? Why can't I buy in realtime, see how it goes and adjust my spend on the fly?

This is starting to happen on the internet, a bit. Search engine marketers certainly do this by integrating planning, buying, and feedback systems. Good, but not good enough.

When I worked at Prodigy (ok, bad example...) we used to track marketing campaigns to determine what the cost of acquiring a customer was compared to the lifetime value of that customer (discounted to the present, of course.) Different marketing campaigns generated customers with different loyalties, different abilities to pay, etc. Memorably, before we strated doing this, Prodigy decided that since their demographic and the demographic of NASCAR overlapped so much, a deal with NASCAR was a really good idea. It wasn't: NASCAR fans did not seem to sign up and if they did, they didn't stay. Bad investment.

Also, memorably, we determined that carpet-bombing the country with sign-up CDs was not a cost-effective marketing technique: the cost of acquiring a customer was some $100, the lifetime value was significantly less. Our explanation for why AOL was doing it? Wall Street was willing to pay $200 per subscriber as a valuation metric. In some variation of the old cliche that in the long run we'll all be owned by a conglomerate, we were right in theory but wrong on company-building. That's the danger of listening to us quants: we're right, but we won't guarantee the time-frame.

What is the dream system? A dashboard that tracks every marketing campaign from spend to lifetime value, gives a real-time ROI for each marketing dollar spent and allows the marketing department to increase or decrease spending on any particular campaign with the push of a button. An ERP system for the marketing department.

Is this doable? In some industries, absolutely. Subscription businesses, where the source of a customer can be linked to a purchase, for instance. A little harder if you're selling $1 cans of fizzy, colored water. But for the vast majority of companies, this type of analysis is probably being done on a spreadsheet kept on the computer of a recent Kellogg grad. That's a little scary.

Tuesday, July 24, 2007

Everything Looks Like a Nail

when you're a hammer. When all you think about is information, everything looks like a problem the internet can solve.

Just read Brad Burnham's post on the sale of Tacoda to AOL/TimeWarner. Before I go into my everything is made of digits spiel, I want to say congratulations to Brad and Fred at USV. I've known them both a long time (and co-invested with Brad when we were both much younger.) I've done business with a lot of venture capitalists and Brad and Fred are stand-outs, not just as businesspeople, but as people (and that, I should add, is just as important.) The sale of Tacoda shows their foresight, if anyone ever doubted it.

Brad makes the point that the internet fundamentally reshapes marketing. I agree. Here's why.

The existence of marketing at all is a bit of an underaddressed topic in economics. Economists don't like marketing because they like markets to be efficient and in an efficient market consumers find the best product for the best price. Marketing doesn't make a product better and it certainly makes it more expensive. Marketing is a sign of an inefficient market, and an especially persistent one.

Plus, marketing is so declasse.

I like to tell my kids that anything they see advertised on TV is absolute junk and I refuse to buy it: the bulk of the purchase price goes to the marketing budget, not the product itself. This makes my life sooo much easier. But it drives them absolutely nuts, because the only products they know of are the ones advertised on TV.

In economics terms, marketing is a way to make finding a product more efficient. Companies advertise so consumers know about their products. In a world of mass media, this cost of search is borne by the seller: it has to be, it would be amazingly inefficient to try and have it be borne directly by the buyer.

What happens when the cost of search goes towards zero? The tables completely turn. Marketing moves (slowly) from being a shotgun fired by the seller to being a tweezer held by the buyer. Behavioral marketing is one of the first steps on that journey and Brad and Fred recognized that years ago. I'm sure they also realize that the really interesting stuff, the revolutionary stuff that makes marketing as we know it today completely obsolete, is what inevitably comes next.

Monday, July 23, 2007

What does non-profit mean?

Not what you think.

Peter Klein over at Organizations and Markets blogged recently about what sound like some really interesting papers on the nature of non-profits. Of course, in a case of the shoemaker's children, I can't find these papers to read, without paying some $30 per (or slogging down to the Business branch of the NYPL in the pouring rain, in defiance of my digital lifestyle.) Ever since my soon-to-be-disabused notion that I would be able to read and understand academic research on quantum computing, I have been getting emailed a daily list of papers being published about it from Arxiv.org. You would think, given their line of work, economists would be hip to the efficiency gains from the free flow of information. But no. I wonder who gets that $30 anyway?

So, instead, I've been reading Evelyn Brody's excellent Agents Without Principals: The Economic Convergence of the Non-Profit and For-Profit Organizational Forms (oh, stop telling me how sexy I am, my head will get big.) She makes an interesting, if obvious in retrospect, point towards the beginning: what distinguishes a non-profit is not profitability but distributions. Non-profits are entities that are not allowed to distribute their profits.

Look at this income statement summary from my alma mater, Columbia University. On $2,501 million of FY2005 revenue, there was some $139 million of EBI (no 'T' for non-profits.) This, of course, has to be so or there wouldn't be an endowment of $5,191 million. The endowment, unless they've been selling equity on the side, has to be retained earnings, or accumulated profit. (OK, I haven't taken the rest of the balance sheet into account, but I really doubt that the endowment is the result of piling up liabilities.)

An EBIT margin of 5.6% is not great, of course. Apollo Group had roughly the same revenue (in 2006), but $650 million of operating income. So, in comparison, Columbia is a non-profit.

Why is Columbia a non-profit? Certainly not because it's not trying to make money (I happen to know that they are very eager to make money; I have the mail to prove it.) Some economists have argued that the inability to distribute profits, and thus enrich the capitalists, leads to more trust from customers when they can not judge the quality of the services the organization is providing: being a non-profit is an antidote to information asymmetry.

If this is true--and it isn't entirely clear that it is--can the information asymmetry be alleviated in some other way, a way that would still allow the organization to be managed by the numbers?

Friday, July 20, 2007

So why not start in two places?

I've been thinking non-stop about internet marketing for twelve years now, so it's easy for me to write about, and I will. What I've been thinking about non-stop for the last six months is entirely different: how can the internet be used to make the world a better place?

I know, it's a stupid question. It's too broad to be interesting. I could use some help, and that's one of the reasons I'm writing.

The first question has to be, what do I mean by better? Peace, Justice, the American Way? Well, sort of. How about: improving the environment, reducing poverty, promoting democracy, and building a more open and tolerant society? Those are the moral criteria Friedman uses in "The Moral Consequences of Economic Growth" and since I like his conclusion--that economic growth can be a force for social betterment--I'm going to use them.

Also, I am generally more interested in the for-profit model than the not-for. The fundamental flaw in most not-for-profits' business model is that (i) they need to raise money to fulfill their mission, and (ii) raising money isn't their mission. They have two separate missions that require two completely different sets of expertise. This is the kind of problem that I used to red-flag when I was a venture capitalist: it's hard enough to become good at one thing, becoming good at two things squared the risk.

In my mind, the ideal pro-social good company makes money by doing good. There is only one mission. It seems to me that the efficiency gains inside the company from spending hard-won dollars on advancing the cause rather than trying to raise still more money are pretty high. I've been talking to or working with several companies over the past few months that do exactly this, and so far I've been blown away by their ideas and execution.

This isn't to say that certain non-profits (Donors Choose is a great example) aren't doing an excellent job leveraging the internet to make the world better. They are. It's just not my focus.

My current conundrum is drawing a heavy theoretical line between creating value by doing good and monetization. That is, the activities I described above as "Better" all create value for people and societies. The best for-profit companies are the ones that create the most value. So, why has the focus on doing good been primarily that of the not-for-profit sector?

Have to start somewhere

I've been staring at this blank blog template for weeks, ever since my friend Brad Burnham told me to stop pestering him with my crazy theories and start blogging about them. That's what the internet is for, isn't it? Peppering the world with your crazy theories?

Procrastination set in.

Then I saw the Hal Varian interview in the Journal. Key quote: "marketing is the new finance." Now that marketers have started to accumulate a vast amount of data, someone needs to make sense of it. More importantly, someone needs to systematize a way of making sense of it.

This isn't a new problem. I remember talking to an exec at Accrue back in '97, he told me the primary feedback he got on the product was that there was too much data for the customer to make sense of. Actually, the problem wasn't even close to new then, either.

The flip side of the coin: the primary cliche of the advertising industry goes back to department store pioneer John Wanamaker: "I know I waste half the money I spend on advertising. The problem is, I don't know which half." More than a hundred years and some $20 trillion (I'm guessing here) of marketing spend later, this is still true, as far as anyone knows. Why is it true? Why is the marketing department the only place in the corporation that doesn't have an automated quantify, report and improve system?

And, more to the point, if half of the $600 billion or so spent on marketing every year is wasted, the companies that can eliminate that waste can earn some piece of the savings.

This, generally, is what I'm going to write about. I don't believe that marketing twenty years from now will resemble today's marketing at all, not even a little bit. With the disclaimer that (i) I sincerely dislike most of today's marketing, and (ii) I'm an optimist, I want to talk about what I think it will look like.