|Shares Out. (in M):||21||P/E||7.9x||7.9x|
|Market Cap (in $M):||323||P/FCF||-||-|
|Net Debt (in $M):||-133||EBIT||0||0|
This isn’t your typical Japanese stock. Aside from the fact that Proto has too much cash and hasn’t yet repurchased shares, the rest of the story is wonderful. It’s a great business (ROEs near 40% without excess cash, 15% with), run for shareholders (founder/Chairman owns 32.5% and company regularly sets ROE goals), growing despite Japan’s well-known demographic headwinds, and it’s available for about 4.3X earnings net of cash (7.2X gross).
[Please note: price and EPS listed above are in Yen. Market cap, net debt, and TEV are in US$]
The story starts with “Goo” which is Japan’s #1 publication/website/mobile app for used car advertising. If you’ve ever seen “AutoTrader” at a gas station here in the U.S., imagine that magazine but with glossy, full color pages. One major difference is that the advertisers are typically used auto dealers, not individuals. (Sales between individuals are not common for used cars in Japan due to the paperwork headaches involved.) Proto has also launched sister publications for motorcycles, imported used cars, and auto parts, which, along with the original Goo are referred to as the “Goo Series”.
Goo and rival “Car-Sensor” are the only two such publications with nationwide coverage – a fact that shouldn’t be too surprising. Whether you’re a seller or a buyer, nationwide coverage is a big plus, and so these two began to dominate smaller, regional rivals even before they touched every part of the country. It wasn’t until just last year that Proto finally reached 100% coverage, after 35 years of building relationships with used car dealers, one at a time. I’m pretty sure we don’t have to worry about new entrants in this market.
…unless they buy Car Sensor, that is. I’m told that Car Sensor is struggling and that its parent – the privately held Recruit Corporation – has put it up for sale. In recent years, Car Sensor went from bi-weekly to weekly publication (while Goo has continued to publish bi-weekly). Its ad rates are typically much lower than those of Goo, while its customers tend to be dealers with lower-end cars. According to Proto, Car Sensor has been undisciplined with pricing and is probably not making much profit, if any.
Some time ago, Proto decided to leverage its relationships with dealers to sell lucrative IT services through a cluster of businesses it refers to as “IT1” in presentations. For example, Proto offers them what are basically outsourced websites, allowing them to display their latest inventory through Proto’s “Goo-Net” website, as well as respond to customer inquiries regarding trade in values and such. Proto sells data on the latest retail prices and successful auction bids for used cars, and offers software that estimates market prices for used cars and forecasts future resale values based on miles driven, etc. Most recently, Proto has begun selling software to manage price estimates to auto maintenance/repair and paint shops. Japan has roughly 90,000 maintenance/repair shops and 45,000 paint shops, while Proto counts only ~5,000 of them as software customers. Thus, the growth opportunity is large.
Proto also has a couple businesses that directly facilitate used car sales, which together are known as “IT2.” “Goo-Net Auction” is a website that enables consumers to put their used car up for auction to dealers, while “Goo-Net Exchange” actually buys used autos for export to markets like Hong Kong and Malaysia.
Finally, Proto has a bunch of fairly unrelated businesses that together don’t amount to much in terms of either revenues or profits. These include the “Lifestyle” segment, which publishes various magazines on the marketplaces for nursing home care, jobs for nurses, et al. Proto also makes some money subletting its own unused real estate, and offers outsourced data entry, performed by its employees who aren’t busy enough inputting data for the Goo Series operations.
As the numbers show, Proto has been fighting Japan’s demographic headwinds and winning:
|New car sales||4,210||4,956||4,609||5,082||5,353||5,739||5,851||5,853||5,828|
|Used car registrations||6,450||6,539||6,698||7,178||7,530||8,066||8,106||8,001||8,195|
There are two factors driving this performance. First, around FY08 (fiscal year ends March 31st), management began to insource database management and fully digitalized the production process for the Goo Series. This led to big savings beginning in FY09. Second, and more importantly, the continual growth in the high-margin IT1 segment has far outpaced growth in any of Proto’s costs.
Specifically, the IT1 segment sports roughly 70% operating margins. Coupled with the high growth, it’s easy to see why this is the most exciting part of the story. At 14%, the Goo Series magazines/websites/apps operating margins are far lower, yet still respectable. Growth has been flattish, although a recent acquisition may change that (more on this below). Importantly, though, the Goo Series is the calling card that Proto uses to get in the door and make the IT1 sale. It’s also the consumer brand that draws in traffic to the Goo-Net website, and consequently makes it possible for Proto to collect and sell better retail pricing data than anyone else.
According to management, the IT2 segment is only nominally profitable. Operating margins in my “other” category (Lifestyle, etc.) have ranged between 3-9% in the past five years. Corporate overhead runs about 1.6B per year.
Proto made its first ever major acquisition in FY12 (consolidated in 3Q12) when it bought a private Malaysian company called MTM for 2.95B Yen (US$ 38M) which is essentially the Malaysian version of Proto’s Goo Series. The price equals about 5X sales, which given the 30% operating margins is 15X EBIT … not an obvious bargain. But according to management, such businesses are rare throughout South East Asia and MTM is basically where Goo was 20 years ago, implying lots of potential. MTM should do 660M revenue and 150M EBIT this year (FY13). The 3 year goal is 1B in revenue and 300M EBIT.
In January, Proto announced its second major deal, this time buying the auto export business of King’s Auto (consolidated 1Q13) for 1.2B Yen. King’s came with about 700M in inventory, so I’d argue the real price was more like 500M, or 7-8X EBIT. Given the need for physical inventory, this business does not excite me. And like the rest of the IT2 segment it has razor thin margins (specifically, 1%). But management says that it helps to cement their relationship with used car dealers, and they think they can get operating margins up to 2-3%.
So maybe these two deals will look really smart in a few years. And maybe they won’t. But in any event, keep in mind that they cost a grand total of 4.15B Yen, or roughly 1.3X a year’s income. Also keep in mind that Proto still has 12.3B in cash. In other words, these aren’t huge acquisitions. It’s not Time Warner buying AOL. It’s quite possible they’ll do another MTM-type acquisition in Singapore and/or Taiwan, but I think these will be small as well. (By the way, a China deal is not in the cards. Management notes that in order for the business model to work, the used car deal market needs to be fragmented as it is in Japan/Malyasia/Singapore/Taiwan, but it’s more consolidated in mainland China.)
Most recently, in early May Proto announced a planned increase in advertising and employment opex of about 1.2B yen in FY13. Given that total FY12 opex had been only 11.4B, this was a big number and the shares promptly tanked. But what management didn’t communicate so well is that 1) the increased advertising cost is non-recurring, and 2) the increased employment costs are largely for revenue-generating positions (sales, etc.). Proto typically ramps up promotion costs around the launches of new business lines (which happens frequently), and history shows the ROI on this has typically been decent. Likewise, I don’t consider it a negative when well-run companies hire to drive profitable growth – even if it might depress results for a couple quarters while new hires get up to speed.
Let’s suppose every last Yen of that opex boost turns out to be a complete waste. The ad spend was a one-time deal which will obviously go away. And salespeople that don’t produce get fired. So FY14 operating income would simply be the FY12 result (5,869M Yen) plus about 200 from MTM and 65 (1% on 6.5B revenue) from King’s Auto. This works out to an FY14 EPS of about 170, which is the number I used to arrive at my 4.3X P/E net of cash.
As it happens, Proto is behind schedule on this hiring program, thus I’m estimating FY13 EPS of roughly 155.
Random final notes:
Proto provides financials and lots more in English: http://www.proto-g.co.jp/IR/english/index.html
IR contact is email@example.com (I’ve found him to be pretty accessible and knowledgeable)
Announced 2-for-1 stock split on 9/3/12