It's a dirty job, but somebody has to do it!

Counting inventory, one item at a time, at retail outlets in the middle of the night might sound like a dirty job, but somebody has to do it.  But that doesn't mean it isn't lucrative - because it is, especially for Ajis, a Japanese listed company (stock code: 4659) that dominates this niche in Japan (market share: nearly 80%) and is replicating that success in other countries like Korea (where they are also the market leader). 

  I think Ajis stock is a buy - excluding treasury shares (from an old stock buyback program), it sells for just 5.5x forward earnings (2.5x ex-cash), 0.95x book value, and 1.4x EV/EBIT.  Ajis' stock price is almost at its all time low, and that is simply too cheap given their track record of 1) averaging 70% pre-tax return on invested capital (defined as EBIT / equity less cash) from FY2005-2010, 2) tripling their sales and quadrupling their EBIT from FY2000-2010, 3) increasing dividends and stock buybacks, and 4) successfully replicating their business model overseas. 

Background:

Ajis (acronym: "All Japan Inventory Services") was founded in 1978 and learned the inventory stocktaking business from an Arcadia, California company called Muscolino Inventory.  Muscolino pioneered this business model in 1940s California and is now owned by Phyle Inventory Services (http://www.picsinv.com/).  

The company had observed early on how big American retailers (i.e. Wal-Mart) had outsourced 100% of their inventory stocktaking and took the view that big Japanese retailers would eventually do the same.  Why?  Because outsourcing inventory stocktaking saves time (so shop staff can focus on customers) and money (it lowers the risk of stealing and increases the accuracy of operational/audit data).  Ajis' thesis was ultimately correct, and today it serves 70 of the top 100 retailers in Japan and is the only company in its industry with a nationwide branch network (90 offices + 4,000 employees).  

Business model:

Ajis' business model is beautifully simple - it's just a bunch of guys in a shop counting food, clothing, hardware, etc. in the middle of the night for a fee per item counted.  This activity is not capital intensive (capex averaged just 2% of sales and from 2005-2010) and is able to generate steady EBIT margins (averaged 10-11% from 2000-2010) because its major cost - people - is actually quite variable since 95% of them are temporary staff who work only according to seasonal demand (i.e. when clients need stuff counted at the end of the quarter).  Just how good is this model?  Ajis has no debt, too much cash, and operates in a high tax jurisdiction like Japan but its return on equity (net income / equity) still averaged 19% from 2005-2010.  

Stock doesn't deserve to be this cheap:

         Ajis stock is trading at a fresh 5 year low (and nearly an all-time low), even though 1) revenue and EBIT in that time increased 50% and 60%, respectively 2) 10% of the shares (0.54 million of 5.4 million total) have been bought back (but not yet cancelled), and 3) dividends per share are up 30%.  What gives?

         To me, Ajis is a classic "fallen angel" story: given the incredible growth it experienced, expectations became unrealistically rosy; the operations (and stock price) were overdue for a hiccup at some point.  That hiccup came in the form of an operational statistic the company is forced to release each month on its website to fulfill JASDAQ exchange listing obligations - the year-on-year percentage change in monthly unconsolidated revenue (see http://www.ajis-group.co.jp/ir/gyouseki.html).  Forget that this information is not very useful to investors in the first place.  The fact is, many short term punters had gotten used to seeing this "comps" number ring positive month after month for almost a decade.  When that number finally flashed (slightly) negative four months in a row, as it did in FY2010 (March-end), the stock price continued a steep descent that began during the late 2008 Lehman shock.

        Yes, Ajis revenue and EBIT declined 2% and 8%, respectively in FY2010 (March-end), and we're still seeing slightly negative monthly comps this year.  But is the business model permanently broken?  No.  If you ask Ajis executives they'll tell you what really happened.

  First, revenues from their stocktaking segment (Y16.8 billion of the Y18.4 billion total revenues) actually declined by less than 1% year on year *.  The EBIT from this segment also declined much less than the 8% decline reported in the consolidated P&L, too.  The same trend was observed in the 1Q of this year (ending June 2010 -- 1Q and 3Q are the seasonal "lows" for counting).   The bulk of the consolidated revenue and EBIT declines both last year and in 1Q really came from their ancillary "temp staffing" business (where they use their expertise in hiring/firing temp staff to place temporary workers in cashier/clerk jobs across the country at various retailers for a fee).  That business was a great fee earner in the past, when finding good cashiers/clerks at the right price was really tough (Ajis was good at this, so retailers would pay Ajis for their expertise). Post-Lehman, however, the supply of desperate cashiers/clerks has increased, so demand for Ajis' temp staff service has declined.     

Second, the average Japanese consumer is still timid and a few struggling retailers (Ajis' clients) got desperate.  These retailers tried 1) closing shops, 2) moving to smaller locations, and even 3) counting inventory internally at a few locations.  But none of these are permanent threats to Ajis' franchise - inventory stocktaking can't be easily replicated by either its competitors or its clients (who theoretically could try to do it in-house). 

Competitors haven't been able to gain share because nationwide retail chains typically contract out entire regions to a single stocktaking company ahead of time.  Ajis, with the only nationwide branch network in the industry, 32 year track record as an industry pioneer with a blue chip client list, and status as the only listed company in its industry (being listed is a big deal in Japan), has a clear advantage in winning/keeping clients.  Where's the proof? In 2003 Ajis had ten competitors - today it has just 2.

Theoretically, Ajis' nationwide retail clients could try counting their own inventory, but the risk/reward ratio wouldn't be favorable.  As an industry the amount of money saved (Ajis' total revenues are just Y18.4 billion) would be peanuts as a percentage of the industry's total revenues (i.e. just one large retailer, Seven & I Holdings, has domestic revenues of Y3.6 trillion!).  Moreover, they'd risk chaos to get it right - a lot of time and aggravation would have to be invested in a non-core activity for a large retailer to build a nationwide stocktaking team.  And as Ajis' track record has shown, in order for a stocktaking team to be profitable (or breakeven) you need 1) many clients (not just one, which is what an in-house stocktaking team would be constrained with) and 2) the ability to handle large numbers of temp staff (hire/fire them on demand, train them quickly, etc.) unrelated to the core retailing business. 

*NOTE: The core stocktaking revenues might have actually increased YoY. Ajis' overseas stocktaking revenues (3% of their total stocktaking segment) were actually hit by an unfavorable exchange rate - the Yen strengthened a lot last year, especially against the Korean won (Ajis' major overseas market) even though the Korean business has been growing briskly in local currency terms.  

Major risk: Technology

In the very long term, the real risk to Ajis' business could be technology - specifically, the use of IC tags and/or RFID technology for organizing inventory.  On July 23, the Wall Street Journal published a story about the increased use of radio tags to track/count clothing at Wal-Mart and other large retailers (article: http://online.wsj.com/article/NA_WSJ_PUB:SB10001424052748704421304575383213061198090.html).  Could this be a serious threat to stocktaking companies in general? Maybe.  Though IC/RFID is currently used mostly for tracing purposes (not counting/checking) and low volume, slow-moving, high price tag items (like jewelry), we can't rule out that one day the technology will be cheap enough and accurate enough to take share from traditional inventory stocktakers - at least for certain goods.  High volume, fast-moving, cheap, and perishable goods are still counted mostly by hand in most developed countries (including the US).

Are such tags going to kill Ajis' franchise in Japan?  No -- the reality is these tags are still far too expensive.  According to the WSJ story above, the sensor tags cost about USD 7-10 cents (or 6-9 Yen) each.  That's cheaper than what they used to be (USD 50 cents, or 43 Yen), but still expensive enough to require Wal-Mart to subsidize their clothing suppliers to accommodate the sensors in their clothing labels. Ajis management believes the tags need to cost far less than 1 Yen each before they'll start sweating.  

Why should we believe them?  First, 72% of Ajis' stocktaking revenue comes from convenience stores, supermarkets, 100 yen (i.e. USD 1 dollar) discount stores, and general merchandise stores - places that sell exactly the kind of high volume, fast-moving, cheap/perishable goods that compete on price.  Adding a 6-9 Yen sensor on top of a 100 Yen price tag is simply too severe a handicap in a country with anemic consumption.   

Second, Seiyu - a Japanese retail subsidiary of Wal-Mart, the same Wal-Mart said to be in love with radio tags in the WSJ story above - is actually a long time Ajis client (!) and according to Ajis' management, Seiyu is NOT changing to sensor tags any time soon.

What's it worth?

         The stock trades at Y1,308 per share. Excluding the 0.54 million shares bought back but still held in treasury (strictly speaking they haven't been cancelled yet so the shares still "exist" but practically they don't - dividends aren't paid to those shares and the company has never sold a treasury share), it has net cash per share of Y716 and book value per share of Y1,373.  Hence the EV per share is Y592 vs. a historical EPS of Y235 and forward EPS (company guidance) of Y238.  Is the company guidance trustworthy?  Yes, Ajis' business model affords them good revenue visibility (clients sign contracts and schedule stocktaking services in advance) and the company has rarely revised down a forecast in its 14 years as a listed company. 

So here we have a high margin, high return on capital business performing a critical value-added service with no debt and 80% market share effectively selling for just 2.5x forward earnings or a 40% earnings yield.  At this price we'd only need to wait 2.5 years to get all our money back in earnings/dividends.  I'll go out on a limb and say this business model will still be healthy in 2.5 years and IC/RFID tags probably won't get cheap enough to take share from it even then. 

If we actually give Ajis some credit for its 32 year old franchise and assign a more realistic (but still conservative) 7-9x P/E to it we'd get to a "core business value" of Y2143 per share.  Add that to its idle cash pile of say, Y670 per share (it's really Y716 per share but let's be conservative), and we get an intrinsic value range of Y2,336-2,812 per share, or 79-115% above the current stock price of Y,1308. 

Catalysts:

1)      Overseas sales/earnings growth, i.e. in Korea where they just became #1 in market share. This portion of their business (China, Taiwan, Korea, and Thailand) is still just 3% of stocktaking revenue but could be meaningful in a few years, especially since many of their clients (i.e. 7-Eleven Japan) have become absolutely dominant in higher-growth places like the Philippines, Thailand, Taiwan, Hong Kong, and Singapore (note: 7-Eleven Japan operates in those countries via independent franchisees, so stocktaking decisions may or may not follow the outsourced pattern seen in Japan).

2)      Recovery of monthly unconsolidated sales comps.  Total retail sales in Japan have increased year on year for 6 consecutive months through June 2010 (http://www.foxbusiness.com/markets/2010/07/28/update-japan-retail-sales-slowdown-likely-ahead/).  Sooner or later the brief negative streak will end and the short term punters who've been dumping it based on this signal will regain interest in the stock.  Also, if the overseas business outside Korea grows further, the monthly sales comps (which includes only the Japanese and Korean stocktaking business) will matter less and less as a bell-weather for the company's performance.

3)      Japan's ratio of outsourced inventory stocktaking across all retail formats is still far less than the United States - especially in categories like mom/pop outfits (much of these exist in Japan versus almost anywhere else) and drug stores (according to research done a few years back by Mizuho, a Japanese bank, some 70% of US drug stores use outsourced stocktaking vs. just 30% in Japan).  Ajis is in a pole position to grab more business here.

4)      Resumption of sell-side analyst coverage - once the market cap dropped below Y10 billion (US$110 million) all brokers started ignoring the stock.

5)      Disappearance of competitors: according to Mizuho Bank, Ajis had 70% market share and 10 competitors in 2003. According to Ajis management, they have 80% market share and just 2 competitors today. Only one of those competitors is 100% focused on inventory stocktaking (Asset Inventory, an unlisted company owned by a private equity-controlled Canadian firm called WIS Inventory).  Ajis' other competitor, a company called P and P, is listed (Japan stock code 2426) but inventory stocktaking is just a minor focus of their overall P&L (which is tiny to begin with).