Novatel Wireless (NVTL)

Flash Player 9 or higher is required to view the chart
Click here to download Flash Player now

View the full NASDAQ:NVTL chart at Wikinvest

Novatel Wireless (NVTL) is a high quality tech company with no debt trading at the value of cash plus investments (enterprise value turns negative at $5.90/share).  Since there’s no chance of the company liquidating and it’s unlikely the company will make any other cash distribution, I can understand why the market is ignoring the cash value.  But the company isn’t a dog.  It’s grown revenue from $104mm in 2004 to $337mm in 2009 while only increasing share count from 29mm to 31mm.  Evidently they’re investing in valuable R&D.  It’s difficult to determine how valuable the R&D is but we’ll try and establish a range by capitalizing R&D and amortizing it at various rates to estimate RoIC.  If NVTL continues to plow earnings into R&D, they might never show a profit, but that doesn’t mean the business won’t grow, in both size and value.  And even if we assume that R&D is effectively a sunk cost, FCF has been significantly positive at $31mm, $17mm, and $34mm in 2009, 2008, and 2007 respectively.  Novatel is profitable and should trade at a higher enterprise value than zero.

The Business

Novatel competes in three primary segments, but the MiFi product is arguably the most important (47% of Q1 sales).  Other segments include USB and data card modems (50% of Q1 sales) and embedded wireless connectivity devices (3% of Q1 sales). 

The MiFi is effectively a cell phone that creates a WiFi bubble around itself, providing up to five nearby devices with internet access.  With very positive initial reviews, a good job by the marketing team on branding, and expected increases in data transmission over wireless networks, initial analyst estimates were optimistic.  But every quarter, sales seem to disappoint.  Initially, it seemed like the dynamics of sell-through vs sell-in were to blame.  Now the upgrade cycle is a problem (expected to resolve in the second half as we transition from 3G to 4G).  But there is one other problem that isn’t being spoken about by analysts.  The MiFi is being sold by 18 wireless carriers worldwide.  You can see which carriers are involved and how geographically diverse they are in the map below:

But it’s difficult to reconcile this map with the fact that 92% of revenue came from North America in Q1.  We can see in the chart below that despite having $33.8mm of revenue from MiFi sales, only $5.6mm of revenue was international:

 

The missing piece to the puzzle is that European telecoms made a marketing error.  In stores where cell phones are typically given away for free or heavily subsidized, the MiFi was sold as a retail product.  Sales never picked up.  Here’s an article discussing the different marketing approaches.  Now, that marketing strategy is being reconsidered.  We can see from the google trends chart below that internet searches for MiFi are strong and likely growing stronger.  This is aided by the fact that every quarter, NVTL announces new carrier partners.

We can also see what country these searches were originated in.  The US and Canada are responsible for a disproportionate share of revenue relative to internet buzz.  I expect some of this buzz to translate into sales in coming quarters as telecoms work out attractive pricing plans.

As difficult as it is to guess what MiFi sales will be, the other products are even worse.  Embedded product (mini card) sales will always be lumpy.  NVTL will constantly have to bid on and win deals in competitive markets to sell mini cards, and the market seems to have all but dried up.  The only guidance we have is an LTE (more advanced technology) design win for a 2011 laptop model with tier-1 laptop manufacturer.  So we know this category won’t go extinct, but it’s hard to estimate what contribution it will make, especially short-term.

The wireless modem business has also been on the decline for a while as competitors have squeezed margins.  Some of this business will fall victim to cannibalization from the MiFi business.  The best I can do is relay management’s sentiment from conference calls.  The story is that we’re late in the 3G technology stage, so NVTL is experiencing normal margin compression, and there should also be some pent-up demand for new products once 4G rollouts start happening in the 2^nd half.  I can’t translate that into a forecast, and analysts who have tired before have failed.

Realizing how difficult it is to estimate future revenue, the best way to value NVTL is to consider what has been spent on R&D in the past, and what sort of a return this R&D has generated.  I realize that it’s the future that matters, not the past.  But the simple fact of the matter is that with at most an 18-month product cycle, NVTL will likely be unrecognizable in three years.  It will be selling different products to different customers.  But we know that NVTL has employees with skills in research, design, marketing, and a management team with a track record of selecting R&D projects.  So if we want to know what the future holds for Novatel, we need to consider its past.

RoR&D

Estimating what sort of return Novatel generates on investment in R&D is a challenge and requires judgement.  This is compounded by the fact that NVTL claims the product cycle has shortened (meaning R&D becomes obsolete faster) from about 18 months three years ago to about 9 months today.  The driver of obsolescence is the product cycle.  Multiple carriers in different countries are operating their networks off two main technologies, EV-DO or HSPA, continually upgrading their networks to provide faster data access.  Competition is intensified as the boundary between phone and TV networks is blurred.  The EV-DO branch of the technology tree is in the transition to WiMax, and HSPA is currently transitioning to the Dual Carrier brand of technology.  The next step in the technological evolution is LTE (short for “long term evolution”) which will be backwards compatible with either EV-DO or HSPA, but should be compatible with each other, improving the economics for R&D.  Half the research will address twice the market.

The future is inherently difficult to predict, but we do have a nice record of what’s happened in the past.  We can use the record to capitalize NVTL’s R&D (move it from an income statement expense to a cash outflow from investing), booking it as an asset, and then depreciating it at some rate.  Readers interested in learning exactly what this entails can read my tutorial here.  Effectively what we’re saying when we expense R&D is that it only benefits the current period.  But R&D benefits more than just the current period, equivalent to the way building a factory to produce widgets benefits future periods.  If we fully depreciated a widget factory in the year we purchased it (ie expensed it), and then spent our earnings every year expanding our widget factory, we would never report earnings, but our profitable widget factory (and hence the value of our company) would continue to grow. 

What makes this analysis complicated is the fact that NVTL does not generate a steady return on R&D.  If they did, we would see a steady RoIC metric when we input the correct depreciation rate.  But we don’t.  The purpose of this analysis isn’t to pinpoint NVTL’s RoR&D, but rather to just get comfortable with the mechanics of the business, and maybe to have a basis for comparison with some other similar companies.  We’ll start by looking at the extreme scenarios and then work our way towards something more reasonable.

Novatel’s financial statements already reflect one extreme end of the R&D capitalization/amortization spectrum, the scenario where all R&D is amortized in one year.  This isn’t totally un-analyzable.  For example, in 2005, NVTL generated $10.5mm EBIT off of $162mm of revenue after spending $20.5mm on R&D.  Because they have no real assets, we have to assume that the only way NVTL was able to generate revenue was from spending money on R&D.  If $20.5mm was the investment and $10.5 was the return, it generated a pre-tax 50% return.  Using this math, we can calculate Novatel’s returns since 2002:

Unfortunately these numbers are lumpy.  Instead what we can do is capitalized R&D starting in 2002, meaning we’ll move the R&D expense line to cash flow from investing, create a balance sheet asset, and then add an “R&D depreciation” line back to the income statement to account for however much of this intangible asset expires each year.  I do not believe the product cycle is (or will be) greater than two years, so we’ll start by expensing 50% of the R&D asset each year.  This generates the following pre-tax RoR&D estimates:

These numbers are highly favorable, but of course they ignore the other more tangible investments made by NVTL.  If we look at an after-tax RoIC number, which takes into account PP&E, goodwill & intangibles, and working capital, we get RoIC numbers that look like this:

 

Not quite as impressive, but still respectable, and deserving of more than zero enterprise value.  Also keep in mind that NVTL only went public in 2001 and in 2002 it generated 1/10 of the revenue it generates today.  So those low return numbers in 2002-2003 should be taken with a grain of salt.

Competition

Novatel competes in a tough environment.  The fact that they lost a significant amount of the traditional embedded business to Qualcomm and Ericsson, and the USB modem business is being eroded by Huawei, is undeniable.  But at the current market valuation, it’s impossible to imagine fierce competition.  If a larger company cared to compete with a product like the MiFi, it would be far easier to just buy NVTL than reproduce the intangible assets. 

Let’s say a company decided they wanted to reproduce the MiFi.  NVTL had R&D expenses of $45mm in 2009, $35 in 2008, and $38 in 2007.  If it cost just $80mm to reproduce the MiFi (R&D for the last two years), a company should be willing to pay $8.82/share ($5.90/share in cash and an $80mm EV).  That would represent a 41% premium to the current share price, and this ignores the value of customer relationships, the value of patents which would have to be worked around (or the cost of litigation), and most importantly the benefit of not having a competitor in the MiFi space.  There is no reason why a company (not even a rinky-dink Chinese shop) would want to compete in this space without simply acquiring NVTL.  After rebates, the company is free.  For investors worried about the difficulty of a takeover, directors and named executive officers own 5.8% of the company and the other 5% holders are giants.  We don’t need an investment banker to talk about synergies to realize Novatel makes an attractive acquisition target.

Conclusion

No matter which way we slice it, it’s difficult to justify the current valuation of Novatel.  For someone with enough money to buy the whole company, it’s free with management included.  For the average stock market investor, we have to pay around $200mm for a company that’s generated average FCF over the past three years of $27mm… still not a bad deal.  This company will likely never report a profit as they plow earnings back into R&D.  But as long as they’re generating a reasonable return on R&D, as they have, the business should continue to grow in size and value.  The value of R&D is evidenced by strong revenue growth.  It might be difficult to value, but not impossible, and with lots of cash, we have solid downside protection. 

Disclosure:  I own shares of NVTL in my kaChing portfolio.