I recommend a short position in GSIC as it has grown its stock price via aggressive and misleading metrics and by overplaying the synergies involved in its three businesses. Wall Street analysts have bought into the company's bogus metrics and the home run synergy story. And management, while receiving a considerable portion of compensation in shares, has been dumping those shares faster than they receive them. And the largest shareholder (Liberty Interactive) has sold all its shares at these levels.

The company consists of three segments. The company started as an e-commerce company serving as the online piece for brick-and-mortar retailers and brands seeking an online presence. They got their start and have their most formidible presence in the sporting goods sector, but they also serve other sectors. This core business is not the greatest business in the world. It has not produced a high ROIC and growth has begun to level off.

In 2007, the company launched its second segment via acquisitions and internal development. This business competes directly against Axciom, Harte-Hanks, Epsilon (part of Alliance Data systems), Experien. etc. The idea was that that GSI Commerce could cross sell marketing services to its e-commerce customers. And this strategy has been fairly successful, but as I will review below, not nearly as successful as the stock price implies.

Finally, in 2009, GSIC launched the third segment via the acquisition of Rue La La (RLL). RLL is a "member-only" shopping site with a bit of multi-level marketing thrown into the mix. There are quite a few of these websites (Gilte Group, Haute Look, Ideeli) as well as similar non-member websites (Overstock). As you can imagine, scoring an invitation to these websites is about as difficult as getting a cup of coffee in the morning. The "members-only" label is purely a marketing ploy. The idea here has again been to cross-sell to its e-commerce customers. This synergy has not yet proven itself, and I am skeptical as to any true synergy in this regard. However, even if moderately successful, the stock remains overpriced.

-

Digging into the numbers takes considerable effort. The company emphasizes two measures that are mis-leading at best.

Shares: 63.3mm * 23.34 = $1447mm

Cash: $64mm

Converts: $150mm face; trade slightly above par; no share dilution in number above

Other Debt: $34mm

NOLs: $150mm (this is probably a bit generous and company has a valuation allowance for most of the NOLs, but they will use them eventually)

Warrants and RSUs: $150mm (not included in share count above)

TEV = $1,597mm

One measure the company emphasizes, Non-GAAP Income from Operations (NGIO), is effectively EBITDA with the caveat that it excludes a very large non-cash compensation expense and repeated acquisition-related expenses.

The company forecasts $135.0mm of NGIO in 2010. However, there is also $30.3mm of stock compensation and $12.8mm of acquisition-related expenses. I assume that the acquisition-related expenses are legitimate one-time expenses, but they seem inordinately large compared to the size of the acquisitions.

Adjusted EBITDA '10 = $104.7mm

Capital Expenditures have slightly exceeded depreciation (not including amortization) so I have simply used Depreciation expense . . .

EBIT '10 = $40.4mm

NOPLAT '10 = $25.9mm (assumes 36% tax rate)

After-Tax Interest = $7mm (note: I include amortization of the converts here because you either need to include amortization of converts or count the value of the optionality of the convert)

Net Income = $19mm

Adjusted EPS = $0.30 (78x)

Rather than this $19mm Adjusted Net Income measure, management has convinced Wall Street to look at its "free cash flow" measure. This measure is skewed because in addition to compensating for one-time expenses, they also add back the stock-compensation expense ($30.3mm) and amortized interest on convert ($10mm) and changes in working capital. Excluding this last adjustment, they will report free cash flow of around $60mm rather than true free cash flow to equity of about $20mm. And this partially explains the mis-perception of the stock.

One thing to note is that the stock compensation expense is a repeated expense that is very real. In 2008, 2009, and 2010, it was $19.4mm, $24.8mm, and $30.3mm, respectively.

Management has effectively convinced Wall Street analysts to play along with the free cash flow measure, announcing in a Feb 10, 2010 press release that: "we achieved free cash flow of $97.2mm, meaningfully exceeding the average analyst expectation of $54.6mm." I will leave it to you to determine how "meaningful" this was.

-

I am going to evaluate the business as a sum of parts. Let me explain why I believe this is proper. Other than the misleading numbers, I believe this is the critical element of the bearish thesis. Management has stated that there are synergies among these three businesses. However, they have offered only two tangible synergies. The first synergy--cross-selling based on e-commerce relationship--is valid and should enable GSIC to grow its marketing services business quicker than otherwise. And it should enable SG&A costs to be spread a bit further. However, this synergy does not justify the tremendous premium over the sum-of-parts value. The second synergy stated by management has been the idea that its e-commerce business gives it a large operational advantage in growing the supply of its RLL business. However, I do not believe there are any synergies to delivering services at a higher quality or cheaper price. The reason is that each of these segments (and the sub-segments of the segments) are intentionally kept discrete by retailers. Fulfillment (the physical inventorying and shipping of goods) is discrete from the website and order processing, which is discrete from customer service call centers. A great example is that 85% of Overstock.com order volume is handled by third-party fulfillment partners. And we know that this number is 100% at eBay. A company can easily (and I would bet there are examples) do its fulfillment with GSIC and sell its overstock through Overstock.com or Gilte Group.

The home run story behind GSIC is that in offering the full suite of e-commerce, marketing, and overstock retail services to customers, they can realize tremendous synergies for a given customer. The problem with this is that at present, many retailers pick and choose each service across vendors and doing it internally. And I think there is a good reason the many vendors in e-commerce services and marketing services have not vertically integrated. If GSIC has stumbled across some winning formula, Axciom, Harte-Hanks, Experien, and Epsilon can quickly add the e-commerce services they do not offer and acquire one of the overstock retailers. However, I believe the natural state of this industry is that the range of e-commerce services (website/ordering, fulfillment, customer service/call center), marketing services, and overstock retail services will always be discrete offerings that can be fit together across vendors and internally.

A more pragmatic reason for a lack of synergy is that the retailers that the e-commerce unit serves are not what brings the shoppers to a site like Rue La La. The predominantly female clientele of Rue La La want fashion apparel, shoes, and purses; home decor; and travel and lifestyle experiences. The e-commerce segment's origins and largest segment is sporting goods. There is limited overlap.

-

The e-commerce services business has always been a mediocre business. GSIC runs the business in two ways. With sporting goods companies, they actually act as the retailer. Pay the retailer some percentage of revenue and use their name and links from website, but they do the merchandising. The second model is the a la carte services, used by all the non-sporting goods companies. This makes revenues a very difficult number to understand because we are comparing apples to oranges.

This segment uses almost all of the Invested Capital, about $170mm. In 2010, this segment produced about $75mm of EBITDA and $30mm of EBIT. Although very difficult to discern because of the apples and oranges revenue, it appears to have grown about 10-15% per year. I do not think anyone would argue this is a great standalone business based on its low ROI (12%-ish), its limited growth, and its lack of a competitive advantage. I have valued it at $400mm, and I think I am being fairly generous.

-

The marketing services segment is a more attractive segment, and even the bears must acknowledge that they have grown this business profitably and quickly. The business competes head-to-head against Harte-Hanks, Acxiom, Epsilon, and Experian, amongst others. The segment will produce $170mm of revenue in 2010 and $13mm of EBIT. Acxiom will produce $1.1bn of revenue in 2010 and has a TEV of about $1.5bn. Harte-Hanks will produce $800mm and has a TEV of about the same amount. Epsilon is tougher to break out, but based on the value of ADS, it is getting an even lower valuation. And note that these businesses all have advantages (more data, more experience, wider offerings) over GSIC. This is inherently a guess, but it is difficult for me to get a valuation over $500mm, which is almost 3x revenue and 38x EBIT. This would entail doubling the business over the next few years and then receiving the same multiple as ACXM.

-

And the final segment (Rue La La) is the wildcard. I would argue that Rue La La is worthless. There is no competitive moat among the top four websites. RLL has not won the eyeball battle. Gilte Group got the quicker start. As stated above, I do not think there are any operational synergies between the e-commerce and marketing segments and the website. I believe the high-end retailers RLL and Gilte, etc. covet will be sure to control their product and move it around the various channels. They are not going to fall captive to a single channel.

This segment has so far not yet produced a profit and this was the main reason for the selloff after the Q2 earnings release. I believe this offers our most significant catalyst. If RLL fails to produce a significant profit in Q3 and Q4, then this becomes a boring services company and the valuation will plunge. GSIC is under pressure to demonstrate that RLL is not another Overstock.com and that because of the other segments, it can add to the bottom line.

-

I think $1bn, or $14 per share is the proper price here.

Risk

- I think the long-term risk here is pretty clear: I am missing the operational synergies that enable GSIC to "own" retailers, leading to a fortuitious cycle of more revenue, higher margins, and deeper moats. Especially on the retail side, this enables them to own the online overstock channel as they get more supply --> leading to more customers --> ad infinitum. I think we will know fairly soon (6 months) whether I have erred.

Q3 and Q4 earnings announcements demonstrating that RLL does not have some special sauce that will enable it to produce profits where others have not produced profits. The home-run story will be off the table. This will lead analysts to bring down 2011 "free cash flow" numbers and earnings as well.