I’m bullish on eBay (EBAY) pretty much for the same reasons why I published Sell recommendations on Amazon.com and Facebook in recent days. It’s all about the combination of growth and sentiment, and the Street’s tendency to mess up quite often when it comes to this approach.

The Logic Behind My Position

For most, including me, I believe objective rules-based strategies are far superior overall to seat-of-the-pants genius. This approach will periodically steer me wrong on individual stocks, but my long experience shows that it points me in the right direction far more often than not. Even where it does go wrong, it tends to do so in ways other than what I anticipate at times when I try to second-guess the model. So like my recent Sell recommendations for AMZN and FB, this Buy recommendation on EBAY is based on probabilities.

The probabilities in both instances involve the investment community’s tendency to make a mess of things when it gets particularly bullish on shares of rapidly growing companies, so much so, that it can pay to simply bet naively on the Street’s ability to misjudge growth stocks. Hence my bearish stances on AMZN and FB. Switching gears, the Buy screen that now calls EBAY to my attention was created as a flip-side version of my earlier Sell screen. I’m now looking for situations in which the Street is likely to also miss the boat when dealing with companies that have had some struggles, as has been the case with EBAY. The model is further described below, after the discussion of EBAY.

A Bullish Case for eBay

EBAY has not been hot lately. In May, 2014, it was plagued by a security breach – an experience which sadly is far from unique in the online world and one that casts dark clouds over businesses for a long time. And on top of that, the online flea-market aura it has traditionally had, while cool in the early days of the internet, is not so wonderful in this generation. Now e-activities are taken for granted and users are more interested in finding what they want to find and doing what they want to do without pausing to experience a sense of circa 2000 newbie amazement. It also doesn't help that by comparison, Amazon.com has been getting bigger and better in eBay’s e-commerce neighborhood. And as if that were weren’t enough, the major search engines, in their quest to continue to provide users with the most relevant search results possible, upgraded their algorithms in ways that wound up giving less prominent placement to eBay’s old-school presentations.

We must, however, maintain some perspective. It’s not as if the company is in trouble.

With 2014 in the rear-view mirror, along with that year’s big deficit-producing write-off, the company’s return on assets over the trailing 12 months was 9.58% versus a 4.84% S&P 500 median. The company has a debt-to-equity ratio of 1.32, versus an S&P 500 median of 0.84. But it can well afford its borrowings, as indicated by its 10.35 interest coverage ratio (versus 7.74 for the S&P 500). Also, eBay’s high return on equity, 27.95% versus 14.28% for the S&P 500, suggests the debt is being deployed productively.

The problem, so to speak, is that good, or even very good, isn’t good enough when investors have been expecting more. And when a company is seen as being in a business similar to that of a firm like Amazon.com, which has been on a tear, investors expect a lot. Sales growth rates that may be more in line with what we think of as struggling brick and mortar (3.64% in the trailing 12 months and a minus 1.26% rate over the past five years, and even the improved 5.62% rate of the latest quarter) absolutely does not cut it for a company like eBay. The P/E ratio is OK at 13.67 times estimated results for the next fiscal year (versus 16.62 for the S&P 500). But the price/sales ratio, 3.65 versus 2.31 for the S&P 500, raises too many eyebrows.

With bullish cases like what I’ve shown to this point, who needs bears!

Actually, the lackluster results can make for a bullish case given the way the Street tends to extrapolate recent trends into the indefinite future, and given that such assumptions often turn out to be incorrect. In last week’s Sell article, we saw tendencies to over-rate future prospects for hot growth companies. Now, we’re dealing with tendencies to under-rate prospects for companies that have hit rough patches. This contrarian type of investment case could be less likely to succeed if company fundamentals were weak. But as noted above, that’s not so with eBay.

We have good reason to expect substantial improvement in sales growth going forward, if not in the next quarter or two, at least down the road. The key, here is the company’s ongoing “structured data” initiative. To get a sense of what this is all about imagine an index card containing the following information:

• John Doe, 1234 First Avenue, Someplace, USA 11111, 999-111-4444

That’s all on one card. It’s fine if you’re content to file everything in one place, alphabetically under “D” for example, and let those who need to find it shuffle from one card to the next until they do. This is an example of what we refer to nowadays as “unstructured data.” With this setup, we’d be able to look up John Doe and find all the information we have on him. But we would not be able to look up folks living in zip code 1111 and find John Doe. Nor could see him on a list of all people named John living in Someplace. Unstructured data is limited in terms of what we can get from it.

This is pretty much how the traditional eBay offering was posted. A seller would type in a bunch of information on the electronic equivalent of an index card. That’s why finding things for sale on eBay had such a hit-and-miss quality. It was fine for the early days of internet, when folks were dazzled by the concepts of anything and everything about life on line. And let’s face it; what eBay had even on day one was a heck of a lot better than index cards.

We’ve come a long way since then. Electronic index cards simply don’t cut it any more. Buyers don’t have the patience to scroll through them. And even worse, when an item is sold, the record is killed, thus reducing the number of available records involving eBay offerings of products in that category, which has the impact of making eBay listings look less important to major search-engine algorithms, thus causing them to decline in positioning.

Enter structured data. Now, imagine a whole bunch of separate index cards like this:

• Card # 1, First Name: John
• Card # 2, Last Name: Doe
• Card # 3, Address – Number: 1234
• Card # 4, Address – Street: First Avenue
• Card # 5, City: Someplace
• Card # 6, State: USA
• Card # 7, Zip Code: 11111
• Card # 8, Area Code: 999
• Card # 9, Phone Number: 111-4444
• Collection of logic A: tells the system that each of these 9 items (cards, or in data-speak, fields) relate to a single entity, or “record.”
• Collection of logic B: tells the system how all of the field #1 entries in all records can relate to one another, how all of the field #2 entries in all records can relate to one another, etc
• Collection of logic C: creates content using all the new database items and the capabilities provided by A and B.
• Phew. That’s a mouthful. Seems like a lot of trouble relative to index cards. But it’s worth the effort and is increasingly becoming the way things are done nowadays. That’s because we can do so much more with this sort of thing. Now, for example, we can create a presentation for zip code 1111 and list, in alphabetical order by last name, everybody who lives there, and drill down to find more on each person.

This is huge for eBay. Now, for example, the company can organize all its offerings, create product pages, enable users to see offerings under each product presented various ways; by user rating, by user review, by price, in the context of feature comparisons, etc. And since the sale of a single item doesn’t kill the product page, it’s easier for it to accumulate the sort of visitation and web “authority” that search-engine algorithms cherish.

(NOTE to readers who are database or SEO professionals: Please don’t laugh. I’m trying.)

This is nothing new. You’ve seen it before in many contexts, including major e-commerce sites. That eBay is implementing it does not put it ahead of the curve. Instead, the implementation enables eBay to remove a major drawback. It’s not a complement that the company is finally doing it now, at long last. But it is an investment issue, specifically, a reason to expect sales and profits to advance much more quickly as the project is completed.

The bad news is that this sort of thing takes a long time. Just compare the simple unstructured John Doe record with the much fancier structured version, and consider how much had to be done to get from the former to the latter. Now imagine what it takes to program all that with a major e-commerce database, and then do things that make use of the new database.

The good news is that eBay is marching steadily forward. At the end of the third quarter, 48% of its offerings had become structured and the company is already creating new pages with what it has so far. At present, though, only a small portion of it is visible to users.

Some might be discouraged by CEO Devin Wenig’s latest conference call comments to the effect that “while we’re moving at an urgent pace, you can expect us to take an intelligent approach to this transition to ensure that we don’t disrupt our customers along the way.” That’s not likely to please the “I want it yesterday” sentiments of many in the investment community. But my time in the dot-com world, including a several-year stint at Reuters during the time when Wenig was C.O.O. there, I know enough to take the phrase “disrupt our customers along the way” very, very seriously. So shareholders should be rooting for eBay to do it right, rather than too-fast.

There are other things going on at eBay. For example, it’s been working on its mobile app, and it has Stub Hub. Ultimately, though, I see the eventual fruits of the structured data initiative as the most likely catalyst for the sort of returns that can make the stock worth buying, or at least worthwhile for those who don’t use stocks as betting chips on whether the company will meet, beat or miss the next quarter’s guidance.

And now, for those who are interested, what follow is more on the model I used.

The Semi-Contrarian Buy Model

The Buy model I created using Portfolio123 that I’m using here is not an exact mirror image of last week’s Sell screen. It would be more accurate to describe it as something that leans generally in the opposite direction. Here are its features:

• Look for run-of-the-mill analyst sentiment (scores below 80 and above 20 in the Portfolio123 Sentiment ranking system I created based on analyst ratings, earnings surprise and estimate revision). In connection with my Sell screen, I showed how analysts tend to be misguided when they get too enthusiastic. On the other hand, I have seen them to be pretty decent at recognizing and steering clear of the worst basket cases.
• Focus on companies with poor scores under my Portfolio123 Growth ranking system, the bottom 20%,
• Add a dash of mundane value, nothing fancy – just a score of 50 or better under my Poretfolio123 Value ranking system (based on plain-vanilla ratios such as P/E, PS and P/B).

Testing this Buy Screen

Figure 1 shows the result of a conventional 10-year backtest of the screen.

Figure 1

(Click on image to enlarge)

Rolling backtests (which examine a lot of independent 3-month holding periods each of which begins on successive Mondays) show that like the Sell screen, this is a better model for up markets. But the balance between up market versus down-market performance is far superior to what we saw last time for the Sell screen.

Table 1: Last 10 Years

Table 2: Last Year

Obviously, anybody would prefer a model that suggested above-market performance under all conditions. But in real life, that can be easier said than done. In this case, what we’re looking for is a potentially decent trade-off between what we might expect in up markets versus down periods, and the screen that produced EBAY, while it can’t assure anything about the future, is at least aligning probabilities in a desirable manner.