Berk Sure Has A Way

I sent my traffic arbitrage post to folks at Yahoo and Google. It looks like the two examples I blogged about were removed from Google – I wonder if they found others.

After my blog post about Marchex’s arbitraging, additional examples were sent to me by another investor. Some of the ads were relevant to the queries below are relevant, others not. One example for “cups online” that takes the user to lensoutlet domain:

Additional examples:

“Endoscopic” has an ad for siliconeimplant.com

“Job Information” has an ad for informationjobs.com

“Flight attendant” has an ad for bigemployment.com

“Atlanta jobs”, ”Boston jobs”, “New York Jobs” has ads for bigemployment.com, sometimes
informationjobs.com (ad on the “more sponsored results” pages)

“Practice GED” has an ad for GED.com

“High School” has an ad for GED.com (ad on the ”more sponsored results” pages)

“Plastic surgery” has an ad for siliconeimplant.com

In addition here is a trick to find more examples on Google partners:

1) Search for one of their domains with the following
format on Google:
+”www.breastenlargementcreams.com” +ads +by
+google +search
2) Click on a cache link next to a result
3) Find the highlight text on the page with the URL
4) Notice the query that caused the ad to be served - many times it is not relevant

Other examples you can use in the same format

“www.siliconeimplant.com”
“www.job-sites.com”
“www.bigemployment.com”
“www.informationJobs.com”
“www.JobOnline.com”
“www.ged.com”
“www.merchantproviders.com”

You will notice that while some of the queries that spawned these ads were somewhat relevant, many are not.

Should Google or Yahoo care about this recycled traffic? After all, YHOO, GOOG, MCHX all win in these situations because they all get paid. Users of Google search and Google’s partner sites lose because they are inconvenienced if they are taken to a place that doesn’t make sense to them – another page of paid links. The real loser may be Yahoo’s advertisers. If the traffic is not relevant, then their high priced ads won’t convert to customers.

I don’t know what Yahoo and Google plan to do, but if I were responsible for it, I would probably be doing a lot of investigating.

From Google’s perspective, I would: 

• Take down the ads as I find them
• Investigate the other ads in same account to see if they are “real” sites and not paid link pages
• Search through all the accounts to find other Marchex domains to see if they have multiple accounts arbitraging at once

From Yahoo’s perspective, I would:

• Investigate the sources of traffic that Marchex provides via web logs, query logs, etc.
• Evaluate conversion tracking trends for Marchex’s pages over time as well as versus other more organic traffic on Yahoo.com for the same keywords
• Check Marchex’s contract and the Yahoo terms of service to see if this is prohibited
• Convene the Marchex account team to determine how to deal with it
• Ask Marchex to outline how widespread this is and to send over their traffic logs that might help fill in the gaps – who knows they might undercover other sketchy traffic sources

Back to the numbers. Who knows how much MCHX has spent arbitraging, but it is clear they are spending more on sales and marketing. Sales and marketing expenses have were $1.32 mil in Q1, $1.54 mil in Q2, and $2.76 mil in Q3. That’s a 79% increase. I wonder where that money is going.

I own puts on MCHX shares at various prices.

Why has Marchex’s stock price not withered like the rest of the tier two search stocks?  This graph shows MCHX, INSP, MIVA, LOOK, and INCX over the past twelve months:

Marchex has increased dramatically while the other players have withered.  Two likely reasons:

1) Smart domain acquisitions have masked deterioration in the rest of their businesses

2) Quarterly acquisitions make it confusing to analyze the stock

The other search players have seen their stocks swoon when the market realizes that they are shrinking, not growing.  If investors get a clearer picture of Marchex’s pro forma revenue numbers, MCHX’s stock price may wither like the rest of the tier two search players.

MCHX Value on a Comparables Basis

Let’s assume that MCHX has not overpaid for acquisitions that the domains and other acquisitions are worth what they paid for them (just this year!) – $164.4 mil for name development, $20.5 mil for Pike Street, $30.6 mil for Industry Brains and $12 mil for new domains for a total of $227.5 million.

Since I see no difference between the rest of MCHX (Enhance / GoClick / TrafficLeader) and the other tier two players, I will use revenue comps (since only INSP is making money).

Applying the aggregate revenue multiple of these comparable tier two players (1.24X) to the TTM revenue of for Enhance/GoClick/TrafficLeader ($62 million) one would arrive at an implied valuation of $76.6 mil for this part of the business.  Add that to $227.5 mil from acquisitions and $5 mil in net cash to get $309 mil valuation.  Divide that by 38.3 mil shares outstanding to get a value of $8.07 per share.

For giggles, if you use INSP as a comp against all of MCHX (the closest full company comparable since INSP owns traffic too), you get a value of $4.35 per share.

Amazingly, Marchex’s enterprise value of $890 mil (using the 10/30/05 close of $23.4), makes MCHX more valuable than INSP, LOOK, MIVA and INCX combined!  Maybe MCHX will go shopping or file a secondary!

I own puts on MCHX shares at various prices.

Marchex bulls would lead you to believe that traffic to their domains is organic and relevant.  I am sure some of it is.

However, it appears that Marchex (or an affiliate) is abusing Google AdWords and exploiting Yahoo by arbitraging traffic.  There are ads on relatively inexpensive words on Google and that direct clickthroughs to unrelated highly monetized Marchex domains. For example, if you do a search for “Arabic” and “measurement” on Google, you will find ads linking to breastenlargementcreams and siliconimplants domains – both owned by Marchex.  Given that almost all the links on these pages are paid with high CPCs –random clickthroughs presumably pump up revenue.

In the screenshots below taken on 11/29/05, see the 4th paid listing on right for “Arabic” and the 2nd and 6th listings for the keyword “measurement”.  Those link to the two Marchex domains with highly monetized Yahoo! keywords.

Hmmm…

I suspect Google would not be pleased with how irrelevant these results are for those terms.  Even more so, I would guess Yahoo! wouldn’t be happy to be sending low quality, irrelevant clicks to their high paying advertisers.  I plan to email some folks at both companies to get their thoughts.

As a Google user and especially as a Yahoo advertiser, I find this very disturbing.  Who knows how much revenue is being generated by these tactics… but it is clearly not sustainable.  It is not organic revenue.  It is likely not even allowed in Google and Yahoo! advertiser and publisher terms of use.

Why would the gem of Marchex’s business need to rely on such sketchy and desperate behavior?  Potential questions:

1. Is the domains business hard to grow? Is this a way to boost revenue on the sly?

2. Is the domains business as profitable as bulls say? How much are they spending on traffic acquisition?

I own puts on MCHX at various prices.

Enhance / GoClick / TrafficLeader still comprise the majority of Marchex’s revenue.  Assuming that TrafficLeader is still a small business (if it is not, that means that Enhance and GoClick are shrinking), it means that Marchex’s biggest business is paid search syndication.  In my first post, I said “Enhance and GoClick (search syndication) are deteriorating low-margin businesses riddled with traffic quality problems and a lack of scale”.

Paid search syndication is a tough business for the non-winners:

1) Network Scale drives Revenue Yield. Search syndication revenue yield is determined by CPCs, breadth of keywords, quality of advertisers and algorithm.  Advertisers are busy and they focus their money and energy on their biggest and best outlets.  A big syndication network leads to more advertisers which leads to 1) higher CPCs, 2) more keyword coverage, 3) more “budget depth” in each keyword covered, 4) better advertiser optimization of campaigns.  Google has over 200,000 advertisers and Yahoo has over 100,000.  Enhance and GoClick do not have scale and consequently do not have high CPCs, broad keyword coverage, budget depth, or frequent advertiser optimization. 

2) Revenue Yield drives Network Scale. In a virtuous cycle, as yields grow higher, the leading search advertising network can attract more scale via new distribution partners. In search syndication, scale drives advantage. Google has decimated the competition in this area.  Yahoo is a decent number two and way ahead of the second tier. Both Google and Yahoo have scale with their own sites and can therefore outbid the second tier for new distribution (MIVA, LOOKD, MCHX, INCX).

3) Traffic sources go with Google (or maybe Yahoo!) – Case in point is Marchex’s direct navigation businesses – Yahoo! provides the paid listings to these sites.   The yield using Google or Yahoo!’s ad network is higher for Marchex even though revenue is split with the network.  Since Enhance and GoClick don’t have the yield of Yahoo!, Marchex’s revenue and profit from these properties would plummet if they made Enhance or GoClick the primary source of paid listings on direct navigation.  Another quick example:  LookSmart.com uses Google paid listings instead of LookListings. 

4) The Golden Traffic Rule – he who has the traffic gets the gold.  Traffic owners get the vast majority of revenue – Google gives out north of 75% of their revenue to their partners.  Yahoo has a hard time signing new quality distribution partners against Google as Google’s yield advantage has grown.  Similarly the only high-profile distribution deals that the second tier have retained or won in the past few years have been low-margin technology platform licensing deals (e.g. MIVA for Lycos, LookSmart for AskJeeves, or Marchex for Yellow page companies) or meta-search deals where more providers improves the value proposition to consumers. Therefore, even if Marchex could compete with Google, Yahoo, etc. to sign up quality distribution partners for Enhance / GoClick – they would end up paying most of the money to partners (plus 3% to Yahoo! for the patent settlement!).

Where is Enhance / GoClick’s traffic coming from?

Some clearly comes from Yahoo!  Marchex’s 10Q’s listed Yahoo! as it largest distribution partner (13% of revenue in Q1, 10% in Q3).  MSN and AskJeeves are also announced partners. What is unknown is how much of the revenue from these partners is related to TrafficLeader customers and how Marchex books revenue for that business line.  Other revenue likely comes from recycled traffic – what I call the rats and mice of the search business – the distribution partners Google and Yahoo didn’t want or rejected.  These distribution partners often include players that utilize pop-unders, pop-ups, adware, spyware, browser-hijacking, gambling-related traffic, international traffic, or just plain fake clicks.  I’ve seen instances where distribution partners have their own distribution partners that have their own distribution partners – creating a kind of traffic laundering that obscures the original source of the click.  Advertisers only want quality traffic – you will notice that all the larger search syndication players talk about traffic quality as an overarching issue.  Now clearly, I don’t know where MCHX’s traffic comes from or how high quality it is, but Overture, Miva and LookSmart all took a hit to revenue and profitability at some point in the name of “cleaning up their traffic”- I suspect MCHX will be no different.

As a starting point, I would encourage all of Enhance and GoClick’s advertisers to use Google’s conversion tracking feature to track the performance of their clicks from Marchex’s distribution providers.  If you have an issue, call Marchex’s customer service line.  If you are thinking about using GoClick or Enhance, you can read what other customers had to say:  Enhance Interactive customer reviews and GoClick customer reviews.

Where about advertisers?

A quick review of some competitive queries suggests Enhance and GoClick do not attract enough advertisers to form a competitive market.  Since Google’s algorithm is more flexible (and therefore higher yielding), we will use Yahoo’s data as a lower-bound proxy:

Oftentimes, second tier search providers will redistribute each other’s listings on their networks. The interesting thing about the rest of the tier two is that they have suffered revenue declines – it will be interesting to see if Marchex can escape the same fate.

What about cross-pollination of advertisers between their products? 

Simply put, I doubt it.

1) Direct Navigation’s revenue share from Yahoo is far superior than MCHX’s yield – MCHX won’t be putting too many (if any) of Enhance or Goclick links on their direct navigation pages – it will hurt revenue and profits

2) TrafficLeader’s larger advertisers would likely bolt if their account rep sent them too much traffic from the lower quality Enhance / GoClick networks.  I have seen local advertisers from their yellow page deals placed (often irrelevantly) within Enhance results from time to time, but on the whole TrafficLeader’s customers are likely sent mainly to the big quality networks. 

3) IndustryBrains has traction with advertisers precisely because they do not distribute beyond the specific sites you buy when you purchase through them.

I own MCHX puts at various prices.

50x cash flow… that’s Google-esque! Few companies are valued greater than 50x cash flow. Those lucky enough generally grow revenue and earnings in the 50%+ range per year for multiple years and boast businesses with very wide moats. MCHX is not one of those businesses.

The math. MCHX has 38.3 million fully diluted common shares outstanding trading at about $22 a share as of today’s close. It has about $55 mil in cash on hand and a liability in the form of convertible preferred shares of about $50 mil. Therefore, MCHX’s enterprise value is about $840 million. MCHX generated $8.6 million in free cash flow so far this year, $2.9 of it in Q3. Trailing 12 month’s cash flow is $9.6 million, implying a ratio of 87. If you annualize the Q3 cash flow to $11.5 million it gives MCHX an EV/FCF ratio of 73. Even if you annualize their peak cash flow quarter of Q2 2005, you still get an EV/FCF ratio of 50. 50! 73? 87!?! Wow.

Growth? As I outlined in my previous post, MCHX’s revenue growth is largely illusory as it has been coming mainly from acquisitions – primarily domain names. Similarly, organic cash flow growth seems to be stuck in neutral or reverse given that MCHX is acquiring new domains with cash flow every quarter this year. Cash flow should be growing at a much greater rate than what they’ve reported:

MCHX fans will say that the direct navigation business is nearly all profit – domains are truly cash flow machines. However, at $7.7 million in direct nav revenue in Q3, if you assume anything above a 40% cash flow margin, the direct nav business provided greater than 100% of MCHX’s cash flow in Q3 2005. This means that cash flow from MCHX’s other businesses must be deteriorating. MCHX only boosted cash flow by $700k over 2004’s Q3 results. Given what they paid for the domains, you would have expected MCHX’s cash flow to be at least $3 – $5 million higher than last year. If you assume the Direct Navigation business gets a 58.5% cash flow margins (90% margin with a 35% tax rate), an interesting tidbit arises. If you compare the first three quarters of 2004 to 2005 for MCHX’s other businesses (excluding direct navigation), reported revenue rose from $28.7 to $48.6 million, while cash flow dropped from positive $1.7 million to negative $1.1 mil. You know you have deteriorating fundamentals when you grow revenue by $20 million but cash flow drops by close to $3 million.

What about next year? Next year’s analyst projections are actually pretty rosy – “earnings” up 47%. But…digging beneath the surface, this makes some sense since their domains businesses were acquired this year making year over year comparisons favorable and “earnings” as tracked by MCHX’s analysts exclude amortization. That means MCHX’s gets to spend shareholder money without it impacting “earnings”. No wonder MCHX keeps buying companies – what a deal!

Wide Moat? MCHX is not the leader in any of its businesses. I will admit the direct navigation business makes good money (for now), but it relies heavily on user error to generate traffic. It strains credibility to think that MCHX will be able to maintain the same level of monetization while making these domains true destinations. As for the rest of MCHX businesses, they are each strategically weak and lack scale (more on that later).

What’s the story? I said that businesses trading at more than 50x cash flow are growing rapidly and have a wide moat. I forgot one case… when a stock is wildly overvalued. That is the story with MCHX’s valuation. Given its risks, MCHX should probably be trading at a EV/FCF of less than 20, implying it should decline 60-75% from $22 before it is back in line with reality.

I own puts on MCHX stock at various prices.

If you dig through Marchex’s financial statements and press releases, you can piece together a picture of how all of Marchex’s businesses would have performed on a pro forma basis. With some conservative estimates for Q4 of last year, it appears that Marchex is not growing organically. Q4, Q1, and Q2 were $24.2, $24.7 and $24.2 million respectively. 

Since Q4 of last year, MCHX managed to grow to $26.3 million in Q3 of 2005 – less than 9% overall and less than 7% over since Q1 2005! 

The numbers look a bit better if you compare year over year with a 25% growth rate. However, since most of that growth likely happened from Q3 to Q4 of 2004, it is clear that the annual growth rate is decelerating rapidly. 

Based on current estimates, they will grow at ~15% Y/Y in Q4. This is not a growth rate of a stock trading at 50 times cash flow. 

See the table below for historical actuals and my estimates of each of their businesses:

Even if you assume they can grow the direct navigation business (more on that later), MCHX’s overall pro-forma revenue will likely be flat in 2006 from 2005. My guess is that they will continue to buy businesses to mask the underperformance of their core business.

Most troubling about these numbers is the fact that all of MCHX cash flow seems to come from the direct navigation (domains) business. That means that investors are valuing MCHX at $800 million on the back of a business MCHX purchased for around $200 million this year.

I own puts on MCHX at various prices.

MCHX is overvalued. I will admit that they’ve made some smart financial moves and they own some profitable internet domain names. However, MCHX shares are probably worth maybe $7, not $20. I don’t know the folks at Marchex and I have nothing against them. But, I am definitely self-interested – I think MCHX will decline so I own MCHX puts at various prices.

I used to work at LookSmart and saw the challenges that tier two search syndication players (MCHX, MIVA, LOOKD, INCX, INSP) face first hand. MCHX has been able to delay a decline through smart acquisitions (paying 10-20x cash flow for domain names funded by selling stock at 50x), but most tier two search syndication business models are unprofitable, strategically challenged, and deteriorating. The key point is that while the other tier two players seem to be more appropriately valued, MCHX is way out of whack. It is important to note that owning end-user traffic combined with a winning tier one syndication business is beautiful, profitable, and often defensible. Google and Yahoo have built great businesses, but MCHX is not in the same league.

Over the course of a few weeks, I will outline my reasons why I think MCHX will decline. In my opinion, it is a combination of the 10 points below. I will try to blog about each of these in turn.

1. Low Organic Revenue Growth – 7% pro-forma revenue growth from Q1 to Q3 that has been masked by acquisitions

2. 50+ P/FCF ratio (on a run-rate basis) is absurdly high 

3. Enhance and GoClick (search syndication) are deteriorating low-margin businesses likely struggling with traffic quality problems and a lack of scale

4. Most tier two search stocks have stumbled (LOOKD, MIVA, INCX, INSP) and the majority of MCHX’s business is very similar to these players

5. MCHX is stretched too thin across too many products with no real synergy

6. TrafficLeader is a former SEM/SEO leader likely struggling with conflicts of interest

7. Direct Navigation is a slow growth business vulnerable to changes in behavior and technology

8. Wacky corporate structure has two tiers of stock and preferred shares. For a company with less than $100 million in revenue?

9. Marchex is pursuing low quality traffic through irrelevant traffic arbitrage.  More details here.
10. Bubble-era accounting does not take into acquisitions into account

I own puts on MCHX stock.

NOTE: I edited this post on 12/6/05 to link to my recent posts. I aslo swapped out reason 9 (MCHX paying too much for acquisitions) with the current 9 (MCHX pursuing sketchy traffic).

J2 Global Communications is my single largest holding and I am still very bullish on their prospects.   

Background

JCOM is an online communications company best known for their flagship product Efax. Efax is the largest provider of web-based faxing (fax-to-email, computer-to-fax). As of the end of Q1 2005, JCOM had 600k subscribed phone lines paying $17 a month. In addition, JCOM has 8.5 million free fax users. Since turning profitable in Q1 2002 JCOM has rapidly grown its earnings. JCOM produced $0.40 a share in Q1 2005.

Valuation 

At $36 a share, JCOM trades forward P/E ratios of 20.6 for 2005 and 16.2 for 2006 earnings estimates. Given that the company has consistently outperformed guidance nearly every quarter for 3 years, the true ratios are likely even lower. The company is estimated to grow at earnings 40%+ this year and next. If you look at the enterprise value (back out cash) divided by run-rate earnings (current quarter earnings * 4), they are trading at one of the lowest valuations since turning profitable (less than 21 run rate earnings). If they hit their guidance when they report on 7/25/05 (which they indicated they would 3 times in June) then that ratio drops to about 18.5. Outside of the huge run up in late 2003, JCOM has mainly traded between 20 and 26 x run rate earnings and their expected growth rate has remained extremely high (40%+). JCOM an enterprise value of about $750 million and a free cash flow run rate of $40 million. Simple math: if you assume that JCOM can get to 3 to 6 million paying subscribers globally, they could be earning 5 to 10 times what they are now.

Positive Business Characteristics

• Telecom is a large market and efax’s millions of free subs indicate people want online faxing
• Consistent and predictable growth
• High margins
• Low customer acquisition costs
• Sticky product (fax number on business card) means low churn – currently close to mobile phone churn rates of sub 3% per month
• International growth prospects (Europe, Japan, China, India, etc.)
• Patent portfolio seems to be preventing large players from challenging JCOM’s position
• Low stock option dilution 

Potential Catalysts – Short Term (1-2 months)

• Could announce stock buyback now that they have over $100 mil in cash
• Raise 2005 guidance on Q2 call
  • Company has raised guidance mid-year in both the past two years
  • International should begin to pick up
• Outperform Q2 guidance
  • Regularly outperforms guidance
  • Just recently started bounty marketing online – which has been hugely successful for other subscription businesses
  • Q2 is generally strongest quarter 

Potential Catalysts – Medium Term (3-6 months)

• Investors acknowledge how undervalued JCOM is
  • EPS growth – still going at 40-50% a year!
  • Run-rate PE is close to lowest points since profitability
    
• Potential for short squeeze
• Major investment banking coverage / magazine coverage 

Potential Catalysts – Long Term (6-18 months)

• January 2006: Set 2006 guidance well above the $2.25 expected (my guess is $2.50 or above)
• Announce a patent licensing deal with a telco or VoIP player
• Larger telco / internet company buys them
• Create a new pricing model to convert free customers (e.g. $20 a year)
• International growth: Launch of a bunch new languages / markets (e.g. Japanese, Chinese)

Risks – I don’t expect these will happen quickly or abruptly

• Competition
  • Large telcos could bundle product
  • Price war by another player
• Advent of online digital signatures
• Gradual decline of fax machines in favor of emails

I was (very) long JCOM at the time of writing.

More precisely, sell all your long term bond funds (10-30 years).  They can and will lose value.  Since bond prices move in the opposite direction to interest rates, when interest rates rise, you will lose money.  And unlike an actual bond, you can’t get your original investment out because funds don’t have an end date.  Luckily, if you sell now (with the 10 year rate at 4%) you did quite well.

Long term interest rates are going higher.  Too many forces working to push them up:

• Fed will raise rates at a measured pace for rest of year – this will shift demand from long to shorter duration bonds
• Budget deficits at record levels with no end in sight
• The likely return of the 30 year long bond
• The eventual halting of Asian central bank purchases (god forbid they sell!)
• The dollar will weaken against Asian currencies sparking a sell off of bonds   
• Inflation looks like it will be coming back (maybe…)

OK, so interest rates rising.  What to do?

• Sell your long term bond funds or ETFs (and junk bond funds too!)
• Buy shorter duration bonds, CDs or a bond fund
• Have a truly long term perspective?  Buy a long bond and hold it until it matures – you get your $1,000 back

• Got sizable assets, want protection and still want decent yield? Build a bond ladder
• Feeling adventurous and have a stomach for risk?  Short TLT (currently $93.57) the iShares 20+ year Bond ETF

I was short TLT at the time of this post.