Cut to the chase,

Yahoo!

(YHOO)

.

It's clear to Yahoo! and any investors who still care about Internet stocks that the company needs to cut its dependence on currently anemic Internet advertising.

So it makes no sense that Yahoo! -- a reasonably smart company that played a crucial role in defining the consumer Internet experience -- hasn't made the one bold move that would give it the stability, independence and growth it craves.

Yahoo! should buy the Internet service provider

EarthLink

(ELNK)

. It would speed up Yahoo!'s recovery. It would open new opportunities for EarthLink. And if I were king, I would make it happen.

Here's Why

Yahoo!'s top priority is marketing services that people will pay for, not just use. So the company is slowly and deliberately adding on for-pay products, ranging from larger email boxes to the forthcoming online music service named pressplay. What better way to prepare people for paying Yahoo! a few bucks a month for music than to get them in the habit of paying Yahoo! $21.95 a month for Internet access? As

AOL Time Warner

(AOL)

figured out long ago, it's easier to get people to open their wallet if you already have their plastic in your pocket.

Those potential future revenues, of course, come on top of EarthLink's current subscription money. It had $319 million in revenue in the third quarter ended Sept. 30 -- nearly half of Yahoo!'s expected sales for all of 2001. If Yahoo! wants to balance advertising money with subscription money, EarthLink is a sensible place to start.

EarthLink would benefit from the deal in two important areas.

One is advertising sales. EarthLink has done a lousy job of selling ads off its 4.8 million subscriber base, although it tries to make the paucity of advertising a badge of honor in its marketing campaigns. Advertising, commerce and content revenues accounted for $4.9 million in the latest quarter, less than 2% of EarthLink's total revenue. Yahoo!'s sales force could do plenty better.

The second payoff for EarthLink is Yahoo!'s brand. Nothing against EarthLink, which has built a loyal subscriber base over the years, but the name comes nowhere near Yahoo!'s among nonsubscribers. That's especially worrisome as EarthLink tries to sign up new customers, either people new to the Internet or those with dial-up connections who are upgrading to high-speed, or broadband, service.

Need Some Scratch

Imagine you're a less-than-sophisticated Internet consumer signing up for broadband Internet. If the choice is between America Online and EarthLink, face it: AOL wins. It just feels safer to take AOL rather than a company you know a lot less about.

But imagine the choice is between AOL and Yahoo!. The competition there is more even. Everyone who has been on the World Wide Web has heard of Yahoo!, and probably used it. And the site has a well-deserved reputation for authoritativeness, reliability and speed -- qualities that reflect well on an Internet service provider.

I can hear the objections to this Yahoo!/EarthLink scheme already. The combination of an Internet portal and an ISP has already created a spectacular failure: the now-bankrupt

Excite@Home

(ATHMQ)

. But this deal would be different. Neither Yahoo! nor EarthLink is burning through cash on the scale of its counterpart in the earlier deal; this would not be a case, to borrow Wall Street's common M&A insult, of two drunks hanging on to each other for support.

Then there's the earnings dilution. The shares that Yahoo! would issue to buy EarthLink, as well as EarthLink's losses, will surely drag down Yahoo!'s earnings per share.

I'm heartened, though, by some rough calculations I made this week on the back of an old Excite@Home stock certificate. Let's say Yahoo! acquires EarthLink in an all-stock deal, paying a 25% premium to EarthLink's share price. Yahoo!, which has about 595 million shares outstanding, would have to issue 231 million more to EarthLink's shareholders. Combine Yahoo!'s expected 2002 profits with EarthLink's losses, stir in the new share count, and you have a company that will earn 5 cents a share in 2002 -- half of what it would earn had I left well enough alone.

But let's try to figure out how we could get the new Yahoo!, a.k.a. Yahoo!Link, back to that 10-cent EPS. Working backward from that dime and the new share count would imply a 2002 profit of $83 million, $42 million more than what you'd get from combining Yahoo!'s and EarthLink's respective expected results.

Where would that $42 million come from? I don't think it's a stretch to say you could squeeze it out of the $1 billion total in sales and marketing, and general and administrative expenses that Yahoo! and EarthLink would spend separately next year. It would mean 4% cutbacks in these expenses, and an

EBITDA margin widening from 6.3% to 8.1%. Those seem like reasonable targets, and they wouldn't touch Yahoo!Link's expected outlays for things like operations and member support or product development.

(For this exercise, I've borrowed J.P. Morgan's separate estimates for Yahoo! and EarthLink. The profits and losses, admittedly, are the quasi-numbers issued by each company to adjust for whatever it wants to adjust for. But since these are what Wall Street pays attention to, I'll work with them.)

Get aggressive in bringing down the combined $775 million cost of revenues -- the two Internet traffic generators have to be able to cut telecommunications expenses somewhere -- and the 10-cent EPS target gets even easier to reach. It won't take an Al Dunlap to handle this job. Any yahoo could do it.